ABS Region Data.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project ABS\n",
    "The goal of this notebook will be to explore data compiled from the Australian Bureau of Statistics (ABS) compiled as part of their [AUSSTATS](https://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/1410.02013-18?OpenDocument) documentation, looking at trends across a wide variety of \"key measures\" by geographical region within Australia.\n",
    "\n",
    "Preliminarily I am interested in exploring the interesting factoids regarding solar panel installations which have piqued my curiousity, and leveraging these for delivering value- creating insights.\n",
    "\n",
    "Note all files were originally \".xls\" format, and have had incredibly minor alterations as follows:\n",
    "* Removing \"branding\" and \"page header\" rows (rows 1-5 in source document)\n",
    "* Removing \"Copyright\" and empty \"footer\" columns at the end of the dataset\n",
    "* Consolidating all the details outlined in the column headers into 1 header per column (removing merged formatting and cross filling for empty cells)\n",
    "* Saving each document as .csv\n",
    "\n",
    "All files are based on the ASGS (Australian Statistical Geography Standard) location methodology. There is potential of further exploration based on documents located [here](https://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/1270.0.55.003July%202011?OpenDocument).\n",
    "\n",
    "### CRISP-DM Process\n",
    "1. Business Understanding\n",
    "2. Data Understanding\n",
    "3. Prepare Data\n",
    "4. Data Modeling\n",
    "5. Evaluate the Results\n",
    "6. Deploy\n",
    "\n",
    "## Main Questions\n",
    "The main questions I would like to explore are:\n",
    "1. What are the main indicators of a region having an increased propensity to have solar panels installed?\n",
    "2. For these identified features, how do changes in the feature impact changes in solar panels installed?\n",
    "3. How have solar installations changed over time when using these features as categories?\n",
    "4. Based on the above, what insights would I share with key stakeholders to the solar industry (energy companies, governments, etc.) when considering future strategies?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Imports Section\n",
    "\n",
    "Note: This section also contains a significant amount of data cleaning and explortation functions I created, feeding into the \"Prepare\" and \"Understand\" steps within the CRISP-DM process."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Declare Imports\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV\n",
    "from sklearn.metrics import make_scorer, r2_score, mean_squared_error\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.ensemble import RandomForestRegressor, AdaBoostRegressor\n",
    "import os\n",
    "from textwrap import wrap\n",
    "import operator\n",
    "\n",
    "solar = 'SOLAR INSTALLATIONS Small-Scale Solar Panel System Installations no.'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set a variable for current notebook's path for various loading/saving mechanisms\n",
    "nb_path = os.getcwd()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create list of all files in CSV directory\n",
    "files = []\n",
    "\n",
    "for (dirpath, dirnames, filenames) in os.walk('{}\\CSV'.format(nb_path)):\n",
    "    files.extend(filenames)\n",
    "    break\n",
    "\n",
    "for f in files:\n",
    "    if f[-4:] != '.csv':\n",
    "        files.remove(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\kempbri\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\IPython\\core\\interactiveshell.py:3020: DtypeWarning: Columns (0) have mixed types. Specify dtype option on import or set low_memory=False.\n",
      "  interactivity=interactivity, compiler=compiler, result=result)\n"
     ]
    }
   ],
   "source": [
    "# Load all the files into a single df\n",
    "df = pd.read_csv('{}\\CSV\\{}'.format(nb_path, files[0]), na_values='-', thousands=',')\n",
    "\n",
    "for file in range(1,len(files)):\n",
    "    df_temp = pd.read_csv('{}\\CSV\\{}'.format(nb_path, files[file]), na_values='-', thousands=',')\n",
    "    df = pd.merge(df, df_temp, how='inner', left_on=['CODE','YEAR','LABEL'], right_on=['CODE','YEAR','LABEL'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>CODE</th>\n",
       "      <th>LABEL</th>\n",
       "      <th>YEAR</th>\n",
       "      <th>NUMBER OF BUSINESSES Number of non-employing businesses no.</th>\n",
       "      <th>NUMBER OF BUSINESSES 1-4 employees no.</th>\n",
       "      <th>NUMBER OF BUSINESSES 5-19 employees no.</th>\n",
       "      <th>NUMBER OF BUSINESSES 20 or more employees no.</th>\n",
       "      <th>NUMBER OF BUSINESSES Total number of businesses no.</th>\n",
       "      <th>BUSINESS ENTRIES Number of non-employing business entries  no.</th>\n",
       "      <th>BUSINESS ENTRIES 1-4 employees no.</th>\n",
       "      <th>...</th>\n",
       "      <th>RELIGIOUS AFFILIATION Judaism   %</th>\n",
       "      <th>RELIGIOUS AFFILIATION Other Religions   %</th>\n",
       "      <th>RELIGIOUS AFFILIATION Secular Beliefs, Other Spiritual Beliefs and No Religious Affiliation  %</th>\n",
       "      <th>RELIGIOUS AFFILIATION Inadequately described or not stated   %</th>\n",
       "      <th>AUSTRALIAN CITIZENSHIP Australian citizen  %</th>\n",
       "      <th>AUSTRALIAN CITIZENSHIP Not an Australian citizen  %</th>\n",
       "      <th>AUSTRALIAN CITIZENSHIP Australian Citizenship - Not stated  %</th>\n",
       "      <th>SPEAKS A LANGUAGE OTHER THAN ENGLISH AT HOME Proportion of total population  %</th>\n",
       "      <th>lat</th>\n",
       "      <th>long</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10102</td>\n",
       "      <td>Queanbeyan</td>\n",
       "      <td>2016</td>\n",
       "      <td>2730.0</td>\n",
       "      <td>1331.0</td>\n",
       "      <td>450.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>4596.0</td>\n",
       "      <td>388.0</td>\n",
       "      <td>173.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.1</td>\n",
       "      <td>1.1</td>\n",
       "      <td>29.2</td>\n",
       "      <td>10.0</td>\n",
       "      <td>87.0</td>\n",
       "      <td>5.4</td>\n",
       "      <td>7.7</td>\n",
       "      <td>12.0</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>149.233725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>101021009</td>\n",
       "      <td>Queanbeyan</td>\n",
       "      <td>2016</td>\n",
       "      <td>402.0</td>\n",
       "      <td>193.0</td>\n",
       "      <td>89.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>712.0</td>\n",
       "      <td>75.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.1</td>\n",
       "      <td>2.1</td>\n",
       "      <td>28.5</td>\n",
       "      <td>10.5</td>\n",
       "      <td>80.6</td>\n",
       "      <td>11.2</td>\n",
       "      <td>8.2</td>\n",
       "      <td>21.2</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>149.233725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10202</td>\n",
       "      <td>Wyong</td>\n",
       "      <td>2016</td>\n",
       "      <td>4573.0</td>\n",
       "      <td>3006.0</td>\n",
       "      <td>895.0</td>\n",
       "      <td>178.0</td>\n",
       "      <td>8657.0</td>\n",
       "      <td>781.0</td>\n",
       "      <td>540.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.4</td>\n",
       "      <td>26.6</td>\n",
       "      <td>8.1</td>\n",
       "      <td>89.7</td>\n",
       "      <td>4.2</td>\n",
       "      <td>6.2</td>\n",
       "      <td>4.8</td>\n",
       "      <td>-33.285641</td>\n",
       "      <td>151.425306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>102021057</td>\n",
       "      <td>Wyong</td>\n",
       "      <td>2016</td>\n",
       "      <td>322.0</td>\n",
       "      <td>193.0</td>\n",
       "      <td>85.0</td>\n",
       "      <td>23.0</td>\n",
       "      <td>631.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>34.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.7</td>\n",
       "      <td>28.6</td>\n",
       "      <td>9.4</td>\n",
       "      <td>86.1</td>\n",
       "      <td>6.3</td>\n",
       "      <td>7.6</td>\n",
       "      <td>8.2</td>\n",
       "      <td>-33.285641</td>\n",
       "      <td>151.425306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10301</td>\n",
       "      <td>Bathurst</td>\n",
       "      <td>2016</td>\n",
       "      <td>2369.0</td>\n",
       "      <td>1123.0</td>\n",
       "      <td>412.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>3991.0</td>\n",
       "      <td>284.0</td>\n",
       "      <td>144.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.4</td>\n",
       "      <td>22.6</td>\n",
       "      <td>11.7</td>\n",
       "      <td>86.9</td>\n",
       "      <td>3.4</td>\n",
       "      <td>9.7</td>\n",
       "      <td>4.0</td>\n",
       "      <td>-33.474707</td>\n",
       "      <td>149.641320</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 589 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        CODE       LABEL  YEAR  \\\n",
       "0      10102  Queanbeyan  2016   \n",
       "1  101021009  Queanbeyan  2016   \n",
       "2      10202       Wyong  2016   \n",
       "3  102021057       Wyong  2016   \n",
       "4      10301    Bathurst  2016   \n",
       "\n",
       "   NUMBER OF BUSINESSES Number of non-employing businesses no.  \\\n",
       "0                                             2730.0             \n",
       "1                                              402.0             \n",
       "2                                             4573.0             \n",
       "3                                              322.0             \n",
       "4                                             2369.0             \n",
       "\n",
       "   NUMBER OF BUSINESSES 1-4 employees no.  \\\n",
       "0                                  1331.0   \n",
       "1                                   193.0   \n",
       "2                                  3006.0   \n",
       "3                                   193.0   \n",
       "4                                  1123.0   \n",
       "\n",
       "   NUMBER OF BUSINESSES 5-19 employees no.  \\\n",
       "0                                    450.0   \n",
       "1                                     89.0   \n",
       "2                                    895.0   \n",
       "3                                     85.0   \n",
       "4                                    412.0   \n",
       "\n",
       "   NUMBER OF BUSINESSES 20 or more employees no.  \\\n",
       "0                                           85.0   \n",
       "1                                           27.0   \n",
       "2                                          178.0   \n",
       "3                                           23.0   \n",
       "4                                           79.0   \n",
       "\n",
       "   NUMBER OF BUSINESSES Total number of businesses no.  \\\n",
       "0                                             4596.0     \n",
       "1                                              712.0     \n",
       "2                                             8657.0     \n",
       "3                                              631.0     \n",
       "4                                             3991.0     \n",
       "\n",
       "   BUSINESS ENTRIES Number of non-employing business entries  no.  \\\n",
       "0                                              388.0                \n",
       "1                                               75.0                \n",
       "2                                              781.0                \n",
       "3                                               40.0                \n",
       "4                                              284.0                \n",
       "\n",
       "   BUSINESS ENTRIES 1-4 employees no.  ...  RELIGIOUS AFFILIATION Judaism   %  \\\n",
       "0                               173.0  ...                                0.1   \n",
       "1                                32.0  ...                                0.1   \n",
       "2                               540.0  ...                                0.1   \n",
       "3                                34.0  ...                                0.1   \n",
       "4                               144.0  ...                                NaN   \n",
       "\n",
       "   RELIGIOUS AFFILIATION Other Religions   %  \\\n",
       "0                                        1.1   \n",
       "1                                        2.1   \n",
       "2                                        0.4   \n",
       "3                                        0.7   \n",
       "4                                        0.4   \n",
       "\n",
       "   RELIGIOUS AFFILIATION Secular Beliefs, Other Spiritual Beliefs and No Religious Affiliation  %  \\\n",
       "0                                               29.2                                                \n",
       "1                                               28.5                                                \n",
       "2                                               26.6                                                \n",
       "3                                               28.6                                                \n",
       "4                                               22.6                                                \n",
       "\n",
       "   RELIGIOUS AFFILIATION Inadequately described or not stated   %  \\\n",
       "0                                               10.0                \n",
       "1                                               10.5                \n",
       "2                                                8.1                \n",
       "3                                                9.4                \n",
       "4                                               11.7                \n",
       "\n",
       "   AUSTRALIAN CITIZENSHIP Australian citizen  %  \\\n",
       "0                                          87.0   \n",
       "1                                          80.6   \n",
       "2                                          89.7   \n",
       "3                                          86.1   \n",
       "4                                          86.9   \n",
       "\n",
       "   AUSTRALIAN CITIZENSHIP Not an Australian citizen  %  \\\n",
       "0                                                5.4     \n",
       "1                                               11.2     \n",
       "2                                                4.2     \n",
       "3                                                6.3     \n",
       "4                                                3.4     \n",
       "\n",
       "   AUSTRALIAN CITIZENSHIP Australian Citizenship - Not stated  %  \\\n",
       "0                                                7.7               \n",
       "1                                                8.2               \n",
       "2                                                6.2               \n",
       "3                                                7.6               \n",
       "4                                                9.7               \n",
       "\n",
       "   SPEAKS A LANGUAGE OTHER THAN ENGLISH AT HOME Proportion of total population  %  \\\n",
       "0                                               12.0                                \n",
       "1                                               21.2                                \n",
       "2                                                4.8                                \n",
       "3                                                8.2                                \n",
       "4                                                4.0                                \n",
       "\n",
       "         lat        long  \n",
       "0 -35.352466  149.233725  \n",
       "1 -35.352466  149.233725  \n",
       "2 -33.285641  151.425306  \n",
       "3 -33.285641  151.425306  \n",
       "4 -33.474707  149.641320  \n",
       "\n",
       "[5 rows x 589 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load in latitude and longitude data\n",
    "df_temp = pd.read_csv('{}\\latlng.csv'.format(nb_path), na_values='-', thousands=',')\n",
    "df_temp.rename(columns = {df_temp.columns[0] : 'LABEL'}, inplace=True)\n",
    "df = pd.merge(df, df_temp, how='inner', left_on='LABEL', right_on='LABEL')\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We are only interested in the very bottom level data (i.e. township, not state). \n",
    "# Given this bottom level has \"CODE\" of 9 characters: remove all rows with code length < 9\n",
    "# Using Len function to evaluate as a string to evaluate because the codes are mixed type.\n",
    "def remove_aggregation(df):\n",
    "    df.drop(df[df['CODE'].astype(str).map(len) <= 8].index, inplace=True)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.reset_index(inplace=True)\n",
    "\n",
    "df = remove_aggregation(df)\n",
    "\n",
    "df.set_index(['CODE', 'LABEL', 'YEAR'], inplace=True)\n",
    "\n",
    "df['Solar Per Household'] = df[solar]/df['HOUSEHOLDS BY TYPE Total households no.']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def clean_data(df, y_column, fill_mean_subset = None):\n",
    "    '''\n",
    "    A function to clean a dataframe and return X & y values for further processing. \n",
    "    Rows are removed where NaNs are present in response vector records.\n",
    "    NaN values for all other features are filled with the mean of the feature.\n",
    "    \n",
    "    INPUT\n",
    "    df - pandas dataframe \n",
    "    y_column - String. Name of column to be used as the response vector\n",
    "    fill_mean_subset - String, column name. Allows the input of a column to \"subset\" when first completing\n",
    "                        imputing missing numerical values with a series mean. E.g. if there is a categorical \n",
    "                        field of \"year\", allows imputing of null values with the mean of each year, rather \n",
    "                        than the mean of the overall series.  \n",
    "    \n",
    "    OUTPUT\n",
    "    X - A matrix holding all of the variables you want to consider when predicting the response\n",
    "    y - the corresponding response vector\n",
    "    '''    \n",
    "    # Remove duplicate columns\n",
    "    drop_cols = []\n",
    "    check_cols = df.columns.tolist()\n",
    "    check_cols.sort()\n",
    "    w_end = len(check_cols)\n",
    "    i = 0\n",
    "    \n",
    "    # Cycle through each column name\n",
    "    while i < w_end:\n",
    "        # assign a Check variable the the column name as a string\n",
    "        # that name string should only include characters up to 1 character after the final space\n",
    "        # e.g. \"* %\" or \"* n\"\n",
    "        check_str = check_cols[i]\n",
    "        check_str = check_str[:(check_str.rfind(\" \")+2)]\n",
    "        \n",
    "        for col in check_cols[(i+1):]:\n",
    "            # look forward in the list of column names for any other items matching CheckString & \"*\"\n",
    "            # add any matches to a list to drop, drop from the \"check\" list as well so make further searches more efficient.\n",
    "            # I'm almost certain there is a more efficient way to do this list/dict-wise\n",
    "            if col.startswith(check_str):\n",
    "                drop_cols.append(col)\n",
    "                check_cols.remove(col)\n",
    "                w_end -= 1\n",
    "        i += 1  \n",
    "    \n",
    "    df.drop(drop_cols, axis = 1, inplace=True)\n",
    "    \n",
    "    # Drop rows with missing response values\n",
    "    df = df.dropna(subset=[y_column], axis=0)\n",
    "    y = df[y_column]\n",
    "    df = df.dropna(how = 'all', axis = 1)\n",
    "    #Drop response column\n",
    "    df = df.drop(y_column, axis=1)\n",
    "    \n",
    "    # Fill numeric columns with the mean    \n",
    "    num_vars = df.select_dtypes(include=['float', 'int']).columns\n",
    "    # First, fill with the mean of the subset based on given category\n",
    "    if fill_mean_subset != None:\n",
    "        index_reset = False\n",
    "        index_names = list(df.index.names)\n",
    "        \n",
    "        # Filtering sucks with multi-indexing so temporarily reset the indexes for this action\n",
    "        if fill_mean_subset in index_names:\n",
    "            index_reset = True\n",
    "            df.reset_index(inplace=True)\n",
    "\n",
    "        #Check if subset variable is an index item\n",
    "        for subset_item in df[fill_mean_subset].unique().tolist():\n",
    "            for col in num_vars:\n",
    "                subset_mean = df[df[fill_mean_subset] == subset_item][col].mean() \n",
    "                df.loc[(df[fill_mean_subset] == subset_item) & (df[col].isnull()), col] = subset_mean\n",
    "        \n",
    "        if index_reset:\n",
    "            df.set_index(index_names, inplace=True)\n",
    "\n",
    "    # For any remaining nulls, fill with the mean of the overall series\n",
    "    for col in num_vars:\n",
    "        df[col].fillna((df[col].mean()), inplace=True)\n",
    "        \n",
    "    # OHE the categorical variables\n",
    "    cat_vars = df.select_dtypes(include=['object']).copy().columns\n",
    "    for var in  cat_vars:\n",
    "        # for each cat add dummy var, drop original column\n",
    "        df = pd.concat([df.drop(var, axis=1), pd.get_dummies(df[var], prefix=var, prefix_sep='_', drop_first=True)], axis=1)\n",
    "    \n",
    "    # Fill OHE NaNs with 0\n",
    "    # Get list of columns after OHE that were not in the \"numeric\" list from earlier, using set function for speed.\n",
    "    cat_vars = list(set(df.columns.tolist()) - set(num_vars.tolist()))\n",
    "    for var in cat_vars:\n",
    "        df[var].fillna(0, inplace=True)\n",
    "    \n",
    "    X = df\n",
    "    return X, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sort_series_abs(S):\n",
    "    'Takes a pandas Series object and returns the series sorted by absolute value'\n",
    "    temp_df = pd.DataFrame(S)\n",
    "    temp_df['abs'] = temp_df.iloc[:,0].abs()\n",
    "    temp_df.sort_values('abs', ascending = False, inplace = True)\n",
    "    return temp_df.iloc[:,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_bar_comparisons(df, num_cols):\n",
    "    '''\n",
    "    Takes a dataframe object and plots each row as a bar chart using the index as the \n",
    "    chart title and the other columns as X-axis labels, within a 5x(n/5) subplot frame\n",
    "    df: dataframe object with values from which you want a chart of each row\n",
    "    num_cols: the number of charts per output row\n",
    "    '''\n",
    "    rows = df.shape[0]\n",
    "    columns = df.shape[1]\n",
    "    fig, axs = plt.subplots(nrows = int(rows/num_cols + rows%num_cols), ncols = num_cols, sharey = True,\n",
    "                           figsize=(15,int(rows*0.66)))\n",
    "    i=1\n",
    "    for row in axs:\n",
    "        for ax in row:\n",
    "            if i <= rows:\n",
    "                sns.barplot(data = df.iloc[i-1:i,:], ax = ax)\n",
    "                ax.title.set_text(df.index[i-1])\n",
    "            else:\n",
    "                ax.axis('off')\n",
    "            i += 1\n",
    "    plt.tight_layout()    \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def feature_plot(importances, X_train, n_features):\n",
    "    \n",
    "    # Display the n most important features\n",
    "    indices = np.argsort(importances)[::-1]\n",
    "    columns = X_train.columns.values[indices[:n_features]]\n",
    "    values = importances[indices][:n_features]\n",
    "    \n",
    "    columns = [ '\\n'.join(wrap(c, 20)) for c in columns ]\n",
    "    \n",
    "    # Create the plot\n",
    "    fig = plt.figure(figsize = (9,n_features))\n",
    "    plt.title(\"Normalized Weights for {} Most Predictive Features\".format(n_features), fontsize = 16)\n",
    "    plt.bar(np.arange(n_features), values, width = 0.6, align=\"center\", color = '#00A000', \n",
    "          label = \"Feature Weight\")\n",
    "    plt.bar(np.arange(n_features) - 0.3, np.cumsum(values), width = 0.2, align = \"center\", color = '#00A0A0', \n",
    "          label = \"Cumulative Feature Weight\")\n",
    "    plt.xticks(np.arange(n_features), columns, rotation = 'vertical')\n",
    "    #plt.xlim((-0.5, 4.5))\n",
    "    plt.ylabel(\"Weight\", fontsize = 12)\n",
    "    #plt.xlabel(\"Feature\", fontsize = 12)\n",
    "    \n",
    "    plt.legend(loc = 'upper right')\n",
    "    plt.tight_layout()\n",
    "    plt.show()  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def feature_plot_h(importances, X_train, n_features):\n",
    "    \n",
    "    # Identify the n most important features\n",
    "    indices = np.argsort(importances)[::-1]\n",
    "    columns = X_train.columns.values[indices[:n_features]]\n",
    "    values = importances[indices][:n_features]\n",
    "    \n",
    "    columns = [ '\\n'.join(wrap(c, 20)) for c in columns ]\n",
    "    \n",
    "    # Create the plot\n",
    "    fig = plt.figure(figsize = (9,n_features))\n",
    "    plt.title(\"Normalized Weights for {} Most Predictive Features\".format(n_features), fontsize = 16)\n",
    "    plt.barh(np.arange(n_features), values, height = 0.6, align=\"center\", color = '#00A000', \n",
    "          label = \"Feature Weight\")\n",
    "    plt.barh(np.arange(n_features) - 0.3, np.cumsum(values), height = 0.2, align = \"center\", color = '#00A0A0', \n",
    "          label = \"Cumulative Feature Weight\")\n",
    "    plt.yticks(np.arange(n_features), columns)\n",
    "    plt.xlabel(\"Weight\", fontsize = 12)\n",
    "    \n",
    "    plt.legend(loc = 'upper right')\n",
    "    \n",
    "    plt.gca().invert_yaxis()\n",
    "    plt.tight_layout()\n",
    "    plt.show()  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "# For fun exploration, let's just run a histogram over everything!\n",
    "def histo_plots(df, num_cols):\n",
    "    '''\n",
    "    Takes a dataframe object and plots each column as a histogram chart using the feature title as the \n",
    "    chart title and the other columns as X-axis labels, within an n*(len/n) subplot frame\n",
    "    df: DataFrame object with values from which you want a chart of each row\n",
    "    num_cols: Integer. the number of charts per output row.\n",
    "    '''\n",
    "    rows = df.shape[0]\n",
    "    columns = df.shape[1]\n",
    "    plot_rows = (int(columns/num_cols) + ((columns%num_cols)!=0))\n",
    "    fig, axs = plt.subplots(nrows = plot_rows, \n",
    "                            ncols = num_cols, sharey = False, figsize=(15,columns))\n",
    "    nlines = 1\n",
    "    \n",
    "    if plot_rows > 1:\n",
    "        for i in range(num_cols * plot_rows):\n",
    "            ax_col = int(i/num_cols)\n",
    "            ax_row = i%num_cols\n",
    "\n",
    "            if i < columns:\n",
    "                try:\n",
    "                    df.iloc[:,i].dropna().plot.hist(ax = axs[ax_col, ax_row])\n",
    "                    axs[ax_col, ax_row].set_title(\"\\n\".join(wrap(str(df.columns[i]), int(100/num_cols))))\n",
    "                    nlines = max(nlines,axs[ax_col, ax_row].get_title().count('\\n'))\n",
    "\n",
    "                except:\n",
    "                    axs[ax_col, ax_row].set_title(\"\\n\".join(wrap(df.columns[i], 100/num_cols)))\n",
    "                    axs[ax_col, ax_row].axis('off')\n",
    "            else:\n",
    "                axs[ax_col, ax_row].axis('off')\n",
    "    else:\n",
    "        for i in range(num_cols):\n",
    "            if i < columns:\n",
    "                try:\n",
    "                    df.iloc[:,i].dropna().plot.hist(ax = axs[i])\n",
    "                    axs[i].set_title(\"\\n\".join(wrap(str(df.columns[i]), int(100/num_cols))))\n",
    "                    nlines = max(nlines,axs[i].get_title().count('\\n'))\n",
    "\n",
    "                except:\n",
    "                    axs[i].set_title(\"\\n\".join(wrap(df.columns[i], 100/num_cols)))\n",
    "                    axs[i].axis('off')\n",
    "            else:\n",
    "                axs[i].axis('off')\n",
    "    \n",
    "    fig.subplots_adjust(hspace=0.5*nlines)\n",
    "    plt.tight_layout()    \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def feature_impact_plot(model, X_train, n_features, y_label):\n",
    "    '''\n",
    "    Takes a trained model and training dataset and synthesises the impacts of the top n features\n",
    "    to show their relationship to the response vector (i.e. how a change in the feature changes\n",
    "    the prediction). Returns n plots showing the variance for min, max, median, 1Q and 3Q.\n",
    "    \n",
    "    INPUTS\n",
    "    model = Trained model in sklearn with  variable \".feature_importances_\". Trained supervised learning model.\n",
    "    X_train = Pandas Dataframe object. Feature set the training was completed using.\n",
    "    n_features = Int. Top n features you would like to plot.\n",
    "    y_label = String. Description of response variable for axis labelling.\n",
    "    '''\n",
    "    # Display the n most important features\n",
    "    indices = np.argsort(model.feature_importances_)[::-1]\n",
    "    columns = X_train.columns.values[indices[:n_features]]\n",
    "    \n",
    "    sim_var = [[]]\n",
    "    \n",
    "    for col in columns:\n",
    "        base_pred = model.predict(X_train)\n",
    "        #add percentiles of base predictions to a df for use in reporting\n",
    "        base_percentiles = [np.percentile(base_pred, pc) for pc in range(0,101,25)]\n",
    "\n",
    "        # Create new predictions based on tweaking the parameter\n",
    "        # copy X, resetting values to align to the base information through different iterations\n",
    "        df_copy = X_train.copy()\n",
    "\n",
    "        for val in np.arange(-X_train[col].std(), X_train[col].std(), X_train[col].std()/50):\n",
    "            df_copy[col] = X_train[col] + val\n",
    "            # Add new predictions based on changed database\n",
    "            predictions = model.predict(df_copy)\n",
    "            \n",
    "            # Add percentiles of these predictions to a df for use in reporting\n",
    "            percentiles = [np.percentile(predictions, pc) for pc in range(0,101,25)]\n",
    "            \n",
    "            # Add variances between percentiles of these predictions and the base prediction to a df for use in reporting\n",
    "            percentiles = list(map(operator.sub, percentiles, base_percentiles))\n",
    "            percentiles = list(map(operator.truediv, percentiles, base_percentiles))\n",
    "            sim_var.append([val, col] + percentiles)\n",
    "\n",
    "    # Create a dataframe based off the arrays created above\n",
    "    df_predictions = pd.DataFrame(sim_var,columns = ['Value','Feature']+[0,25,50,75,100])\n",
    "    \n",
    "    # Create a subplot object based on the number of features\n",
    "    num_cols = 2\n",
    "    subplot_rows = int(n_features/num_cols) + int(n_features%num_cols)\n",
    "    fig, axs = plt.subplots(nrows = subplot_rows, ncols = num_cols, sharey = True, figsize=(15,5*subplot_rows))\n",
    "\n",
    "    nlines = 1\n",
    "\n",
    "    # Plot the feature variance impacts\n",
    "    for i in range(axs.shape[0]*axs.shape[1]):\n",
    "        if i < len(columns):\n",
    "            # Cycle through each plot object in the axs array and plot the appropriate lines\n",
    "            ax_row = int(i/num_cols)\n",
    "            ax_column = int(i%num_cols)\n",
    "            \n",
    "            axs[ax_row, ax_column].plot(df_predictions[df_predictions['Feature'] == columns[i]]['Value'],\n",
    "                     df_predictions[df_predictions['Feature'] == columns[i]][50])\n",
    "            \n",
    "            axs[ax_row, ax_column].set_title(\"\\n\".join(wrap(columns[i], int(100/num_cols))))\n",
    "            \n",
    "            # Create spacing between charts if chart titles happen to be really long.\n",
    "            nlines = max(nlines, axs[ax_row, ax_column].get_title().count('\\n'))\n",
    "\n",
    "            axs[ax_row, ax_column].set_xlabel('Simulated +/- change to feature'.format(y_label))\n",
    "            \n",
    "            # Format the y-axis as %\n",
    "            if ax_column == 0:\n",
    "                vals = axs[ax_row, ax_column].get_yticks()\n",
    "                axs[ax_row, ax_column].set_yticklabels(['{:,.2%}'.format(x) for x in vals])\n",
    "                axs[ax_row, ax_column].set_ylabel('% change to {}'.format(y_label))\n",
    "        \n",
    "        # If there is a \"spare\" plot, hide the axis so it simply shows ans an empty space\n",
    "        else:\n",
    "            axs[int(i/num_cols),int(i%num_cols)].axis('off')\n",
    "    \n",
    "    # Apply spacing between subplots in case of very big headers\n",
    "    fig.subplots_adjust(hspace=0.5*nlines)\n",
    "    \n",
    "    # Return the plot\n",
    "    plt.tight_layout()    \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Start of exploration - \"Data Understanding\" Phase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import the files for exploration\n",
    "# Note: could optimize to import all the CSVs from that directory, not sure how to assign distinct dataframes \n",
    "# for each off the top of my head.\n",
    "df_income = pd.read_csv('{}\\CSV\\Income_ASGS_Final.csv'.format(nb_path), na_values='-', thousands=',')\n",
    "df_pop = pd.read_csv('{}\\CSV\\Population and People_ASGS.csv'.format(nb_path), na_values='-', thousands=',')\n",
    "df_solar = pd.read_csv('{}\\CSV\\Land and Environment_ASGS.csv'.format(nb_path), na_values='-', thousands=',')\n",
    "df_fam = pd.read_csv('{}\\CSV\\Family and Community_ASGS.csv'.format(nb_path), na_values='-', thousands=',')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "keep_cols = ['CODE',\n",
    "             'LABEL',\n",
    "             'YEAR',\n",
    "             'SOLAR INSTALLATIONS Small-Scale Solar Panel System Installations no.']\n",
    "\n",
    "drop_cols = []\n",
    "\n",
    "for column in df_solar.columns:\n",
    "    if not(column in keep_cols):\n",
    "        drop_cols.append(column)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_solar.drop(drop_cols,axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# For the purposes of this analysis, I'm only going to be looking at the most granular data.\n",
    "# This is in order to get the greatest number of unique data points for any ML to be applied.\n",
    "# Therefore we can drop any records with \"CODE\" length < 9. \n",
    "\n",
    "df_list = [df_income, df_solar, df_fam, df_pop]\n",
    "df_income, df_solar, df_fam, df_pop = [df.pipe(removeaggregation) for df in df_list]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>CODE</th>\n",
       "      <th>LABEL</th>\n",
       "      <th>YEAR</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Employee income earners no.</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Employee income earners - median age  years</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Total Employee income $</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Median Employee income $</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Mean Employee income $</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Employee income as main source of income %</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Own unincorporated business income earners no.</th>\n",
       "      <th>...</th>\n",
       "      <th>COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 10 and above kms</th>\n",
       "      <th>COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 9 and below kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Postgraduate Degree kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Graduate Diploma and Graduate Cert kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Bachelor Degree kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Advanced Diploma and Diploma kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate III &amp; IV kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate I &amp; II kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 10 and above kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 9 and below kms</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>Australia</td>\n",
       "      <td>2011</td>\n",
       "      <td>10541561.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>5.280639e+11</td>\n",
       "      <td>41181.0</td>\n",
       "      <td>50094.0</td>\n",
       "      <td>77.3</td>\n",
       "      <td>1809787.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>Australia</td>\n",
       "      <td>2014</td>\n",
       "      <td>11508888.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>6.488015e+11</td>\n",
       "      <td>45725.0</td>\n",
       "      <td>56374.0</td>\n",
       "      <td>77.5</td>\n",
       "      <td>1850832.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>Australia</td>\n",
       "      <td>2015</td>\n",
       "      <td>12035235.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>6.981709e+11</td>\n",
       "      <td>46999.0</td>\n",
       "      <td>58011.0</td>\n",
       "      <td>76.8</td>\n",
       "      <td>1906521.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>Australia</td>\n",
       "      <td>2016</td>\n",
       "      <td>12171794.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>7.249347e+11</td>\n",
       "      <td>48413.0</td>\n",
       "      <td>59559.0</td>\n",
       "      <td>76.7</td>\n",
       "      <td>1958072.0</td>\n",
       "      <td>...</td>\n",
       "      <td>15.4</td>\n",
       "      <td>15.2</td>\n",
       "      <td>10.4</td>\n",
       "      <td>10.1</td>\n",
       "      <td>10.1</td>\n",
       "      <td>10.6</td>\n",
       "      <td>10.8</td>\n",
       "      <td>8.9</td>\n",
       "      <td>9.1</td>\n",
       "      <td>7.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>New South Wales</td>\n",
       "      <td>2011</td>\n",
       "      <td>3292675.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>1.713970e+11</td>\n",
       "      <td>41727.0</td>\n",
       "      <td>52054.0</td>\n",
       "      <td>77.1</td>\n",
       "      <td>578700.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 287 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  CODE            LABEL  YEAR  \\\n",
       "0    0        Australia  2011   \n",
       "1    0        Australia  2014   \n",
       "2    0        Australia  2015   \n",
       "3    0        Australia  2016   \n",
       "4    1  New South Wales  2011   \n",
       "\n",
       "   ESTIMATES OF PERSONAL INCOME Employee income earners no.  \\\n",
       "0                                         10541561.0          \n",
       "1                                         11508888.0          \n",
       "2                                         12035235.0          \n",
       "3                                         12171794.0          \n",
       "4                                          3292675.0          \n",
       "\n",
       "   ESTIMATES OF PERSONAL INCOME Employee income earners - median age  years  \\\n",
       "0                                               38.0                          \n",
       "1                                               38.0                          \n",
       "2                                               38.0                          \n",
       "3                                               38.0                          \n",
       "4                                               38.0                          \n",
       "\n",
       "   ESTIMATES OF PERSONAL INCOME Total Employee income $  \\\n",
       "0                                       5.280639e+11      \n",
       "1                                       6.488015e+11      \n",
       "2                                       6.981709e+11      \n",
       "3                                       7.249347e+11      \n",
       "4                                       1.713970e+11      \n",
       "\n",
       "   ESTIMATES OF PERSONAL INCOME Median Employee income $  \\\n",
       "0                                            41181.0       \n",
       "1                                            45725.0       \n",
       "2                                            46999.0       \n",
       "3                                            48413.0       \n",
       "4                                            41727.0       \n",
       "\n",
       "   ESTIMATES OF PERSONAL INCOME Mean Employee income $  \\\n",
       "0                                            50094.0     \n",
       "1                                            56374.0     \n",
       "2                                            58011.0     \n",
       "3                                            59559.0     \n",
       "4                                            52054.0     \n",
       "\n",
       "   ESTIMATES OF PERSONAL INCOME Employee income as main source of income %  \\\n",
       "0                                               77.3                         \n",
       "1                                               77.5                         \n",
       "2                                               76.8                         \n",
       "3                                               76.7                         \n",
       "4                                               77.1                         \n",
       "\n",
       "   ESTIMATES OF PERSONAL INCOME Own unincorporated business income earners no.  \\\n",
       "0                                          1809787.0                             \n",
       "1                                          1850832.0                             \n",
       "2                                          1906521.0                             \n",
       "3                                          1958072.0                             \n",
       "4                                           578700.0                             \n",
       "\n",
       "                                                 ...                                                 \\\n",
       "0                                                ...                                                  \n",
       "1                                                ...                                                  \n",
       "2                                                ...                                                  \n",
       "3                                                ...                                                  \n",
       "4                                                ...                                                  \n",
       "\n",
       "   COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 10 and above kms  \\\n",
       "0                                                NaN                                                    \n",
       "1                                                NaN                                                    \n",
       "2                                                NaN                                                    \n",
       "3                                               15.4                                                    \n",
       "4                                                NaN                                                    \n",
       "\n",
       "   COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 9 and below kms  \\\n",
       "0                                                NaN                                                   \n",
       "1                                                NaN                                                   \n",
       "2                                                NaN                                                   \n",
       "3                                               15.2                                                   \n",
       "4                                                NaN                                                   \n",
       "\n",
       "   COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Postgraduate Degree kms  \\\n",
       "0                                                NaN                                                     \n",
       "1                                                NaN                                                     \n",
       "2                                                NaN                                                     \n",
       "3                                               10.4                                                     \n",
       "4                                                NaN                                                     \n",
       "\n",
       "   COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Graduate Diploma and Graduate Cert kms  \\\n",
       "0                                                NaN                                                                    \n",
       "1                                                NaN                                                                    \n",
       "2                                                NaN                                                                    \n",
       "3                                               10.1                                                                    \n",
       "4                                                NaN                                                                    \n",
       "\n",
       "   COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Bachelor Degree kms  \\\n",
       "0                                                NaN                                                 \n",
       "1                                                NaN                                                 \n",
       "2                                                NaN                                                 \n",
       "3                                               10.1                                                 \n",
       "4                                                NaN                                                 \n",
       "\n",
       "   COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Advanced Diploma and Diploma kms  \\\n",
       "0                                                NaN                                                              \n",
       "1                                                NaN                                                              \n",
       "2                                                NaN                                                              \n",
       "3                                               10.6                                                              \n",
       "4                                                NaN                                                              \n",
       "\n",
       "   COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate III & IV kms  \\\n",
       "0                                                NaN                                                      \n",
       "1                                                NaN                                                      \n",
       "2                                                NaN                                                      \n",
       "3                                               10.8                                                      \n",
       "4                                                NaN                                                      \n",
       "\n",
       "   COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate I & II kms  \\\n",
       "0                                                NaN                                                    \n",
       "1                                                NaN                                                    \n",
       "2                                                NaN                                                    \n",
       "3                                                8.9                                                    \n",
       "4                                                NaN                                                    \n",
       "\n",
       "   COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 10 and above kms  \\\n",
       "0                                                NaN                                                   \n",
       "1                                                NaN                                                   \n",
       "2                                                NaN                                                   \n",
       "3                                                9.1                                                   \n",
       "4                                                NaN                                                   \n",
       "\n",
       "   COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 9 and below kms  \n",
       "0                                                NaN                                                 \n",
       "1                                                NaN                                                 \n",
       "2                                                NaN                                                 \n",
       "3                                                7.8                                                 \n",
       "4                                                NaN                                                 \n",
       "\n",
       "[5 rows x 287 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Merge the dataframes together for analysis\n",
    "df = pd.merge(df_income, df_solar, how='inner', left_on=['CODE','YEAR','LABEL'], right_on=['CODE','YEAR','LABEL'])\n",
    "\n",
    "cols_to_use = df_pop.columns.difference(df.columns).tolist()\n",
    "cols_to_use.extend(['CODE','YEAR','LABEL'])\n",
    "\n",
    "df = pd.merge(df, df_pop[cols_to_use], how='inner', left_on=['CODE','YEAR','LABEL'], right_on=['CODE','YEAR','LABEL'])\n",
    "\n",
    "cols_to_use = df_fam.columns.difference(df.columns).tolist()\n",
    "cols_to_use.extend(['CODE','YEAR','LABEL'])\n",
    "\n",
    "df = pd.merge(df, df_fam, how='inner', left_on=['CODE','YEAR','LABEL'], right_on=['CODE','YEAR','LABEL'])\n",
    "\n",
    "df.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Employee income earners no.</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Employee income earners - median age  years</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Total Employee income $</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Median Employee income $</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Mean Employee income $</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Employee income as main source of income %</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Own unincorporated business income earners no.</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Own unincorporated business income earners - median age  years</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Total Own unincorporated business income $</th>\n",
       "      <th>ESTIMATES OF PERSONAL INCOME Median Own unincorporated business income $</th>\n",
       "      <th>...</th>\n",
       "      <th>COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 10 and above kms</th>\n",
       "      <th>COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 9 and below kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Postgraduate Degree kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Graduate Diploma and Graduate Cert kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Bachelor Degree kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Advanced Diploma and Diploma kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate III &amp; IV kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate I &amp; II kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 10 and above kms</th>\n",
       "      <th>COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 9 and below kms</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CODE</th>\n",
       "      <th>LABEL</th>\n",
       "      <th>YEAR</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">0</th>\n",
       "      <th rowspan=\"4\" valign=\"top\">Australia</th>\n",
       "      <th>2011</th>\n",
       "      <td>10541561.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>5.280639e+11</td>\n",
       "      <td>41181.0</td>\n",
       "      <td>50094.0</td>\n",
       "      <td>77.3</td>\n",
       "      <td>1809787.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>4.164409e+10</td>\n",
       "      <td>10064.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2014</th>\n",
       "      <td>11508888.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>6.488015e+11</td>\n",
       "      <td>45725.0</td>\n",
       "      <td>56374.0</td>\n",
       "      <td>77.5</td>\n",
       "      <td>1850832.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>4.528178e+10</td>\n",
       "      <td>10953.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015</th>\n",
       "      <td>12035235.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>6.981709e+11</td>\n",
       "      <td>46999.0</td>\n",
       "      <td>58011.0</td>\n",
       "      <td>76.8</td>\n",
       "      <td>1906521.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>4.901755e+10</td>\n",
       "      <td>11581.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016</th>\n",
       "      <td>12171794.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>7.249347e+11</td>\n",
       "      <td>48413.0</td>\n",
       "      <td>59559.0</td>\n",
       "      <td>76.7</td>\n",
       "      <td>1958072.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>5.055164e+10</td>\n",
       "      <td>11669.0</td>\n",
       "      <td>...</td>\n",
       "      <td>15.4</td>\n",
       "      <td>15.2</td>\n",
       "      <td>10.4</td>\n",
       "      <td>10.1</td>\n",
       "      <td>10.1</td>\n",
       "      <td>10.6</td>\n",
       "      <td>10.8</td>\n",
       "      <td>8.9</td>\n",
       "      <td>9.1</td>\n",
       "      <td>7.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <th>New South Wales</th>\n",
       "      <th>2011</th>\n",
       "      <td>3292675.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>1.713970e+11</td>\n",
       "      <td>41727.0</td>\n",
       "      <td>52054.0</td>\n",
       "      <td>77.1</td>\n",
       "      <td>578700.0</td>\n",
       "      <td>47.0</td>\n",
       "      <td>1.390790e+10</td>\n",
       "      <td>10551.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 284 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                           ESTIMATES OF PERSONAL INCOME Employee income earners no.  \\\n",
       "CODE LABEL           YEAR                                                             \n",
       "0    Australia       2011                                         10541561.0          \n",
       "                     2014                                         11508888.0          \n",
       "                     2015                                         12035235.0          \n",
       "                     2016                                         12171794.0          \n",
       "1    New South Wales 2011                                          3292675.0          \n",
       "\n",
       "                           ESTIMATES OF PERSONAL INCOME Employee income earners - median age  years  \\\n",
       "CODE LABEL           YEAR                                                                             \n",
       "0    Australia       2011                                               38.0                          \n",
       "                     2014                                               38.0                          \n",
       "                     2015                                               38.0                          \n",
       "                     2016                                               38.0                          \n",
       "1    New South Wales 2011                                               38.0                          \n",
       "\n",
       "                           ESTIMATES OF PERSONAL INCOME Total Employee income $  \\\n",
       "CODE LABEL           YEAR                                                         \n",
       "0    Australia       2011                                       5.280639e+11      \n",
       "                     2014                                       6.488015e+11      \n",
       "                     2015                                       6.981709e+11      \n",
       "                     2016                                       7.249347e+11      \n",
       "1    New South Wales 2011                                       1.713970e+11      \n",
       "\n",
       "                           ESTIMATES OF PERSONAL INCOME Median Employee income $  \\\n",
       "CODE LABEL           YEAR                                                          \n",
       "0    Australia       2011                                            41181.0       \n",
       "                     2014                                            45725.0       \n",
       "                     2015                                            46999.0       \n",
       "                     2016                                            48413.0       \n",
       "1    New South Wales 2011                                            41727.0       \n",
       "\n",
       "                           ESTIMATES OF PERSONAL INCOME Mean Employee income $  \\\n",
       "CODE LABEL           YEAR                                                        \n",
       "0    Australia       2011                                            50094.0     \n",
       "                     2014                                            56374.0     \n",
       "                     2015                                            58011.0     \n",
       "                     2016                                            59559.0     \n",
       "1    New South Wales 2011                                            52054.0     \n",
       "\n",
       "                           ESTIMATES OF PERSONAL INCOME Employee income as main source of income %  \\\n",
       "CODE LABEL           YEAR                                                                            \n",
       "0    Australia       2011                                               77.3                         \n",
       "                     2014                                               77.5                         \n",
       "                     2015                                               76.8                         \n",
       "                     2016                                               76.7                         \n",
       "1    New South Wales 2011                                               77.1                         \n",
       "\n",
       "                           ESTIMATES OF PERSONAL INCOME Own unincorporated business income earners no.  \\\n",
       "CODE LABEL           YEAR                                                                                \n",
       "0    Australia       2011                                          1809787.0                             \n",
       "                     2014                                          1850832.0                             \n",
       "                     2015                                          1906521.0                             \n",
       "                     2016                                          1958072.0                             \n",
       "1    New South Wales 2011                                           578700.0                             \n",
       "\n",
       "                           ESTIMATES OF PERSONAL INCOME Own unincorporated business income earners - median age  years  \\\n",
       "CODE LABEL           YEAR                                                                                                \n",
       "0    Australia       2011                                               47.0                                             \n",
       "                     2014                                               47.0                                             \n",
       "                     2015                                               47.0                                             \n",
       "                     2016                                               47.0                                             \n",
       "1    New South Wales 2011                                               47.0                                             \n",
       "\n",
       "                           ESTIMATES OF PERSONAL INCOME Total Own unincorporated business income $  \\\n",
       "CODE LABEL           YEAR                                                                            \n",
       "0    Australia       2011                                       4.164409e+10                         \n",
       "                     2014                                       4.528178e+10                         \n",
       "                     2015                                       4.901755e+10                         \n",
       "                     2016                                       5.055164e+10                         \n",
       "1    New South Wales 2011                                       1.390790e+10                         \n",
       "\n",
       "                           ESTIMATES OF PERSONAL INCOME Median Own unincorporated business income $  \\\n",
       "CODE LABEL           YEAR                                                                             \n",
       "0    Australia       2011                                            10064.0                          \n",
       "                     2014                                            10953.0                          \n",
       "                     2015                                            11581.0                          \n",
       "                     2016                                            11669.0                          \n",
       "1    New South Wales 2011                                            10551.0                          \n",
       "\n",
       "                                                                         ...                                                 \\\n",
       "CODE LABEL           YEAR                                                ...                                                  \n",
       "0    Australia       2011                                                ...                                                  \n",
       "                     2014                                                ...                                                  \n",
       "                     2015                                                ...                                                  \n",
       "                     2016                                                ...                                                  \n",
       "1    New South Wales 2011                                                ...                                                  \n",
       "\n",
       "                           COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 10 and above kms  \\\n",
       "CODE LABEL           YEAR                                                                                                       \n",
       "0    Australia       2011                                                NaN                                                    \n",
       "                     2014                                                NaN                                                    \n",
       "                     2015                                                NaN                                                    \n",
       "                     2016                                               15.4                                                    \n",
       "1    New South Wales 2011                                                NaN                                                    \n",
       "\n",
       "                           COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 9 and below kms  \\\n",
       "CODE LABEL           YEAR                                                                                                      \n",
       "0    Australia       2011                                                NaN                                                   \n",
       "                     2014                                                NaN                                                   \n",
       "                     2015                                                NaN                                                   \n",
       "                     2016                                               15.2                                                   \n",
       "1    New South Wales 2011                                                NaN                                                   \n",
       "\n",
       "                           COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Postgraduate Degree kms  \\\n",
       "CODE LABEL           YEAR                                                                                                        \n",
       "0    Australia       2011                                                NaN                                                     \n",
       "                     2014                                                NaN                                                     \n",
       "                     2015                                                NaN                                                     \n",
       "                     2016                                               10.4                                                     \n",
       "1    New South Wales 2011                                                NaN                                                     \n",
       "\n",
       "                           COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Graduate Diploma and Graduate Cert kms  \\\n",
       "CODE LABEL           YEAR                                                                                                                       \n",
       "0    Australia       2011                                                NaN                                                                    \n",
       "                     2014                                                NaN                                                                    \n",
       "                     2015                                                NaN                                                                    \n",
       "                     2016                                               10.1                                                                    \n",
       "1    New South Wales 2011                                                NaN                                                                    \n",
       "\n",
       "                           COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Bachelor Degree kms  \\\n",
       "CODE LABEL           YEAR                                                                                                    \n",
       "0    Australia       2011                                                NaN                                                 \n",
       "                     2014                                                NaN                                                 \n",
       "                     2015                                                NaN                                                 \n",
       "                     2016                                               10.1                                                 \n",
       "1    New South Wales 2011                                                NaN                                                 \n",
       "\n",
       "                           COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Advanced Diploma and Diploma kms  \\\n",
       "CODE LABEL           YEAR                                                                                                                 \n",
       "0    Australia       2011                                                NaN                                                              \n",
       "                     2014                                                NaN                                                              \n",
       "                     2015                                                NaN                                                              \n",
       "                     2016                                               10.6                                                              \n",
       "1    New South Wales 2011                                                NaN                                                              \n",
       "\n",
       "                           COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate III & IV kms  \\\n",
       "CODE LABEL           YEAR                                                                                                         \n",
       "0    Australia       2011                                                NaN                                                      \n",
       "                     2014                                                NaN                                                      \n",
       "                     2015                                                NaN                                                      \n",
       "                     2016                                               10.8                                                      \n",
       "1    New South Wales 2011                                                NaN                                                      \n",
       "\n",
       "                           COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate I & II kms  \\\n",
       "CODE LABEL           YEAR                                                                                                       \n",
       "0    Australia       2011                                                NaN                                                    \n",
       "                     2014                                                NaN                                                    \n",
       "                     2015                                                NaN                                                    \n",
       "                     2016                                                8.9                                                    \n",
       "1    New South Wales 2011                                                NaN                                                    \n",
       "\n",
       "                           COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 10 and above kms  \\\n",
       "CODE LABEL           YEAR                                                                                                      \n",
       "0    Australia       2011                                                NaN                                                   \n",
       "                     2014                                                NaN                                                   \n",
       "                     2015                                                NaN                                                   \n",
       "                     2016                                                9.1                                                   \n",
       "1    New South Wales 2011                                                NaN                                                   \n",
       "\n",
       "                           COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 9 and below kms  \n",
       "CODE LABEL           YEAR                                                                                                    \n",
       "0    Australia       2011                                                NaN                                                 \n",
       "                     2014                                                NaN                                                 \n",
       "                     2015                                                NaN                                                 \n",
       "                     2016                                                7.8                                                 \n",
       "1    New South Wales 2011                                                NaN                                                 \n",
       "\n",
       "[5 rows x 284 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Add in multi-index\n",
    "df.set_index(['CODE', 'LABEL', 'YEAR'], inplace=True)\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['ESTIMATES OF PERSONAL INCOME Employee income earners no.',\n",
       "       'ESTIMATES OF PERSONAL INCOME Employee income earners - median age  years',\n",
       "       'ESTIMATES OF PERSONAL INCOME Total Employee income $',\n",
       "       'ESTIMATES OF PERSONAL INCOME Median Employee income $',\n",
       "       'ESTIMATES OF PERSONAL INCOME Mean Employee income $',\n",
       "       'ESTIMATES OF PERSONAL INCOME Employee income as main source of income %',\n",
       "       'ESTIMATES OF PERSONAL INCOME Own unincorporated business income earners no.',\n",
       "       'ESTIMATES OF PERSONAL INCOME Own unincorporated business income earners - median age  years',\n",
       "       'ESTIMATES OF PERSONAL INCOME Total Own unincorporated business income $',\n",
       "       'ESTIMATES OF PERSONAL INCOME Median Own unincorporated business income $',\n",
       "       ...\n",
       "       'COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 10 and above kms',\n",
       "       'COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Year 9 and below kms',\n",
       "       'COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Postgraduate Degree kms',\n",
       "       'COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Graduate Diploma and Graduate Cert kms',\n",
       "       'COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Bachelor Degree kms',\n",
       "       'COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Advanced Diploma and Diploma kms',\n",
       "       'COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate III & IV kms',\n",
       "       'COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Certificate I & II kms',\n",
       "       'COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 10 and above kms',\n",
       "       'COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Year 9 and below kms'],\n",
       "      dtype='object', length=284)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "#histo_plots(df.iloc[:,4:],5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Age Pension - DVA no.                                              0.988433\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Parenting Payment - Single no.                                     0.987579\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Service Pension - DVA no.                                          0.986307\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Youth Allowance (Other) no.                                        0.986187\n",
       "DWELLING STRUCTURE Separate house  no.                                                                       0.985872\n",
       "SOCIAL MARITAL STATUS & REGISTERED MARITAL STATUS Married in a de facto marriage no.                         0.985244\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Newstart Allowance no.                                             0.984872\n",
       "HOUSING SUITABILITY Dwellings with bedrooms spare no.                                                        0.984345\n",
       "FAMILIES BY TYPE One parent families, children under 15 &/or dependent students no.                          0.983930\n",
       "METHOD OF TRAVEL TO WORK - EMPLOYED PERSONS Other - Employed but did not go to work no.                      0.983315\n",
       "FAMILIES BY TYPE Couple families without children no.                                                        0.982754\n",
       "METHOD OF TRAVEL TO WORK - EMPLOYED PERSONS Used one method - Other (inc. Taxis) no.                         0.982413\n",
       "METHOD OF TRAVEL TO WORK - EMPLOYED PERSONS Used one method - Car (as driver or passenger) no.               0.982220\n",
       "HOUSEHOLDS BY TYPE Lone person households no.                                                                0.979739\n",
       "ESTIMATES OF PERSONAL INCOME Superannuation and annuity income earners no.                                   0.979697\n",
       "METHOD OF TRAVEL TO WORK - EMPLOYED PERSONS Other - Worked from home no.                                     0.979397\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  65-69 years no.                             0.978871\n",
       "HOUSEHOLDS BY TYPE Total households no.                                                                      0.978776\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Age Pension - Centrelink no.                                       0.978695\n",
       "METHOD OF TRAVEL TO WORK - EMPLOYED PERSONS Used one method - Motor bike/scooter no.                         0.978571\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  65-69 years no..2                           0.978430\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  70-74 years no.                             0.978404\n",
       "HOUSEHOLDS BY TYPE Family households no.                                                                     0.978334\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  65-69 years no..1                           0.977937\n",
       "FAMILIES BY TYPE Total families no.                                                                          0.977823\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  10-14 years no..1                           0.977803\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  70-74 years no..2                           0.977791\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Disability Support Pension no.                                     0.977746\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  10-14 years no..2                           0.977681\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  60-64 years no.                             0.977569\n",
       "                                                                                                               ...   \n",
       "TOTAL PERSONAL INCOME (WEEKLY) - PERSONS OVER 15 YEARS Persons earning $1-$499 per week %                    0.003237\n",
       "COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Managers kms                   -0.003204\n",
       "COUNT OF ALL CHILDREN IN FAMILY Average number of all children in a family no.                              -0.003125\n",
       "ESTIMATES OF PERSONAL INCOME Mean Investment income $                                                        0.002632\n",
       "TOTAL PERSONAL INCOME (WEEKLY) - PERSONS OVER 15 YEARS Persons with a negative income %                     -0.002566\n",
       "TENURE TYPE Rented %                                                                                        -0.001855\n",
       "ESTIMATES OF PERSONAL INCOME Median Employee income $                                                       -0.001800\n",
       "ESTIMATES OF PERSONAL INCOME Mean Total income (excl. Government pensions) $                                -0.001372\n",
       "RELIGIOUS AFFILIATION Buddhism   %                                                                          -0.001318\n",
       "UNPAID CHILD CARE  Persons who cared for other child/children  %                                             0.001316\n",
       "COMMUTING TO WORK Average commuting distance from Place of Usual Residence - Males kms                      -0.001262\n",
       "COMMUTING TO WORK Median commuting distance from Place of Usual Residence - Postgraduate Degree kms         -0.000997\n",
       "ESTIMATES OF PERSONAL INCOME Total income (excl. Government pensions and allowances) - Second Quartile %    -0.000895\n",
       "COMMUTING TO WORK Average commuting distance to place of work kms                                            0.000880\n",
       "ESTIMATES OF PERSONAL INCOME Mean Employee income $                                                         -0.000877\n",
       "HOUSEHOLD STRESS Households where rent payments are less than 30% of household income %                     -0.000872\n",
       "ESTIMATES OF PERSONAL INCOME Median Total income (excl. Government pensions) $                              -0.000850\n",
       "UNPAID CHILD CARE  Persons who did not provide child care (%) %                                              0.000731\n",
       "OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Born in Southern & Eastern Europe  %              -0.000649\n",
       "HOUSEHOLD STRESS Households with rent payments greater than or equal to 30% of household income %            0.000280\n",
       "ESTIMATES OF PERSONAL INCOME Total income (excl. Government pensions and allowances) - Highest Quartile %    0.000168\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Carer Allowance no.                                                     NaN\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Parenting Payment - Partnered no.                                       NaN\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Commonwealth Rent Assistance no.                                        NaN\n",
       "REGIONAL MIGRATION  Departures no.                                                                                NaN\n",
       "REGIONAL MIGRATION  Departures no..1                                                                              NaN\n",
       "REGIONAL MIGRATION  Net Migration no.                                                                             NaN\n",
       "REGIONAL MIGRATION  Net Migration no..1                                                                           NaN\n",
       "REGIONAL MIGRATION Internal Migration Arrivals no.                                                                NaN\n",
       "REGIONAL MIGRATION Overseas Migration Arrivals no.                                                                NaN\n",
       "Name: SOLAR INSTALLATIONS Small-Scale Solar Panel System Installations no., Length: 283, dtype: float64"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Investigate correlations to check out items which stand out as potential drivers\n",
    "sort_series_abs(df.dropna(subset=[solar]).corr().loc[:,solar])[1:50]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Initial Thoughts\n",
    "Very interesting that there is a high correlation between number of individuals on government pensions and those earning annuities, and even moreso that there is a *negative* correlation with the share of high income earners and number of solar panels, which also flows through into total household income.\n",
    "\n",
    "Obviously the first step I'm missing is *per household* solar panel information, as simply increasing the number of households in a given region also increases the chance there will be another solar panel installation. \n",
    "\n",
    "I'm also anticipating there will be some outlier suburbs etc. where there is zero probability of household solar panel installation, simply because there are no households (i.e. those comprised entirely of apartment blocks, if such a region exists).\n",
    "\n",
    "I also wonder if there is a reason to explore *change* in installations year on year."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CODE       LABEL                                    YEAR\n",
      "0          Australia                                2011    7760322.0\n",
      "                                                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016    8286077.0\n",
      "1          New South Wales                          2011    2471299.0\n",
      "                                                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016    2604314.0\n",
      "2          Victoria                                 2011    1944690.0\n",
      "                                                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016    2112706.0\n",
      "3          Queensland                               2011    1547304.0\n",
      "                                                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016    1656831.0\n",
      "4          South Australia                          2011     619041.0\n",
      "                                                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016     638792.0\n",
      "5          Western Australia                        2011     794157.0\n",
      "                                                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016     866777.0\n",
      "6          Tasmania                                 2011     192826.0\n",
      "                                                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016     197575.0\n",
      "7          Northern Territory                       2011      60893.0\n",
      "                                                    2014          NaN\n",
      "                                                              ...    \n",
      "128011603  Caringbah South                          2015          NaN\n",
      "                                                    2016       4024.0\n",
      "128011604  Cronulla - Kurnell - Bundeena            2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016       9527.0\n",
      "128011605  Lilli Pilli - Port Hacking - Dolans Bay  2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016        967.0\n",
      "128011606  Woolaware - Burraneer                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016       2583.0\n",
      "128021533  Heathcote - Waterfall                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016       2203.0\n",
      "128021534  Illawong - Alfords Point                 2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016       3158.0\n",
      "128021535  Menai - Lucas Heights - Woronora         2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016       6586.0\n",
      "128021536  Oyster Bay - Como - Jannali              2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016       6957.0\n",
      "128021537  Royal National Park                      2016          NaN\n",
      "128021538  Sutherland - Kirrawee                    2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016       8135.0\n",
      "128021607  Engadine                                 2014          NaN\n",
      "                                                    2015          NaN\n",
      "                                                    2016       5645.0\n",
      "Name: HOUSEHOLDS BY TYPE Total households no., Length: 3114, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print(df['HOUSEHOLDS BY TYPE Total households no.'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['Solar Per Household'] = df[solar]/df['HOUSEHOLDS BY TYPE Total households no.']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SOLAR INSTALLATIONS Small-Scale Solar Panel System Installations no.                              1.000000\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Age Pension - DVA no.                                   0.988433\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Parenting Payment - Single no.                          0.987579\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Service Pension - DVA no.                               0.986307\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Youth Allowance (Other) no.                             0.986187\n",
       "DWELLING STRUCTURE Separate house  no.                                                            0.985872\n",
       "SOCIAL MARITAL STATUS & REGISTERED MARITAL STATUS Married in a de facto marriage no.              0.985244\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Newstart Allowance no.                                  0.984872\n",
       "HOUSING SUITABILITY Dwellings with bedrooms spare no.                                             0.984345\n",
       "FAMILIES BY TYPE One parent families, children under 15 &/or dependent students no.               0.983930\n",
       "METHOD OF TRAVEL TO WORK - EMPLOYED PERSONS Other - Employed but did not go to work no.           0.983315\n",
       "FAMILIES BY TYPE Couple families without children no.                                             0.982754\n",
       "METHOD OF TRAVEL TO WORK - EMPLOYED PERSONS Used one method - Other (inc. Taxis) no.              0.982413\n",
       "METHOD OF TRAVEL TO WORK - EMPLOYED PERSONS Used one method - Car (as driver or passenger) no.    0.982220\n",
       "HOUSEHOLDS BY TYPE Lone person households no.                                                     0.979739\n",
       "ESTIMATES OF PERSONAL INCOME Superannuation and annuity income earners no.                        0.979697\n",
       "METHOD OF TRAVEL TO WORK - EMPLOYED PERSONS Other - Worked from home no.                          0.979397\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  65-69 years no.                  0.978871\n",
       "HOUSEHOLDS BY TYPE Total households no.                                                           0.978776\n",
       "SELECTED GOVERNMENT PENSIONS & ALLOWANCES Age Pension - Centrelink no.                            0.978695\n",
       "Name: SOLAR INSTALLATIONS Small-Scale Solar Panel System Installations no., dtype: float64"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sort_series_abs(df.dropna(subset=[solar]\n",
    "                         ).corr().loc[:,solar])[:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Solar Per Household                                                                                                                                  1.000000\n",
       "RENT & MORTGAGE PAYMENTS Average monthly household mortgage payment $                                                                               -0.588355\n",
       "RENT & MORTGAGE PAYMENTS Average monthly household rental payment $                                                                                 -0.534415\n",
       "TOTAL PERSONAL INCOME (WEEKLY) - PERSONS OVER 15 YEARS Persons earning $500-$999 per week %                                                          0.510632\n",
       "GIFTS/DONATIONS REPORTED BY TAXPAYERS Median $                                                                                                      -0.509600\n",
       "SOCIO-ECONOMIC INDEXES FOR AREAS (SEIFA) Australia SEIFA decile ranking - Index of Education and Occupation (IEO) no.                               -0.493421\n",
       "SOCIO-ECONOMIC INDEXES FOR AREAS (SEIFA) Australia SEIFA decile ranking - Index of Relative Socio-Economic Advantage and Disadvantage (IRSAD) no.   -0.491016\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  25-34 years %                                                                      -0.487650\n",
       "OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Born in Americas  %                                                                       -0.486215\n",
       "SOCIO-ECONOMIC INDEXES FOR AREAS (SEIFA) State SEIFA decile ranking - Index of Education and Occupation (IEO) no.                                   -0.483511\n",
       "EQUIVALISED TOTAL HOUSEHOLD INCOME - FAMILY, LONE PERSON AND GROUP HOUSEHOLDS Median equivalised total household income (weekly) $                  -0.477242\n",
       "SOCIO-ECONOMIC INDEXES FOR AREAS (SEIFA) State SEIFA decile ranking - Index of Relative Socio-Economic Advantage and Disadvantage (IRSAD) no.       -0.471988\n",
       "ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  55-64 years %                                                                       0.462778\n",
       "SPEAKS A LANGUAGE OTHER THAN ENGLISH AT HOME Proportion of total population  %                                                                      -0.460249\n",
       "ESTIMATES OF PERSONAL INCOME Investment income earners - median age years                                                                            0.455646\n",
       "ESTIMATES OF PERSONAL INCOME Mean Employee income $                                                                                                 -0.454605\n",
       "OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Total Born Overseas  %                                                                    -0.453626\n",
       "OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Born in Southern & Eastern Europe  %                                                      -0.447117\n",
       "TOTAL PERSONAL INCOME (WEEKLY) - PERSONS OVER 15 YEARS Persons earning $2000-$2999 per week %                                                       -0.446976\n",
       "ESTIMATES OF PERSONAL INCOME Mean Total income (excl. Government pensions) $                                                                        -0.445860\n",
       "Name: Solar Per Household, dtype: float64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sort_series_abs(df.dropna(subset=['Solar Per Household']\n",
    "                         ).corr().loc[:,'Solar Per Household'])[:20]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Next level thoughts\n",
    "Okay, so the data at least makes sense - there is a strong negative correlation between population density and solar installations per household, owing to some key factors regarding building composition within the region:\n",
    "* More renters in the region result in less solar investment (they are unlikely to reap the long term capital benefits)\n",
    "* Greater number of apartment buildings\n",
    "* I'm not sure about this one, but number of households with neither rooms to spare nor rooms needed, so perfectly utilised. My best guess is this figure itself is closely associated with renting.\n",
    "\n",
    "Some other observations leap out at me:\n",
    "* The positive correlation with 55-64 year old residents\n",
    "* The negative correlation with immigrants and non-citizens\n",
    "* The negative correlation with overall income, but positive correlation with income between $500-999 per week\n",
    "* Positive correlations with income earners' ages\n",
    "* A *very* surprising negative correlation with gifts and donations! Goes to show that this particular \"green\" signal is not an indicator of altruistic behaviour.\n",
    "\n",
    "It is also interesting to note the overwhelming amount of strong negative correlations, solar has a lot going against it apparently!\n",
    "\n",
    "Annoyingly, there seem to be some direct duplicate columns in the dataset (note the \"no..1 and no..2\" variables) so I will need some way of eliminating these before running PCA and other ML techniques over the data to ensure that coefficients aren't distorted by this duplication.\n",
    "\n",
    "Finally, it looks like if I want to measure solar per household in order to identify drivers, I will be limited to using the 2016 dataset. As identified above in some data exploration, the \"Total Households\" number is only valid for census years, and solar panels data was only recorded after 2014.\n",
    "\n",
    "# But that's just correlations\n",
    "Time to delve into some more advanced modelling to see what jumps out!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# \"Data Modelling\" - Random Forest\n",
    "\n",
    "First we'll start out doing a gridsearch to find a random forest model that best fits all the data, and understand the most important features from that in order to answer questions 1 & 2:\n",
    "> 1. What are the main indicators of a region having an increased propensity to have solar panels installed?\n",
    "> 2. For these identified features, how do changes in the feature impact changes in solar panels installed?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Use the function to create X and y\n",
    "X, y = clean_data(df.drop(solar, axis=1), 'Solar Per Household', 'CODE')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split the 'features' and 'income' data into training and testing sets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 12 candidates, totalling 36 fits\n",
      "[CV] min_samples_leaf=1, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=20, total=   0.8s\n",
      "[CV] min_samples_leaf=1, n_estimators=20 .............................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.8s remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .............. min_samples_leaf=1, n_estimators=20, total=   0.8s\n",
      "[CV] min_samples_leaf=1, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=20, total=   0.8s\n",
      "[CV] min_samples_leaf=1, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=40, total=   1.5s\n",
      "[CV] min_samples_leaf=1, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=40, total=   1.9s\n",
      "[CV] min_samples_leaf=1, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=40, total=   1.8s\n",
      "[CV] min_samples_leaf=1, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=80, total=   3.5s\n",
      "[CV] min_samples_leaf=1, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=80, total=   3.3s\n",
      "[CV] min_samples_leaf=1, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=80, total=   3.3s\n",
      "[CV] min_samples_leaf=1, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=1, n_estimators=160, total=   6.9s\n",
      "[CV] min_samples_leaf=1, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=1, n_estimators=160, total=   8.2s\n",
      "[CV] min_samples_leaf=1, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=1, n_estimators=160, total=   9.8s\n",
      "[CV] min_samples_leaf=2, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=20, total=   0.7s\n",
      "[CV] min_samples_leaf=2, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=20, total=   0.8s\n",
      "[CV] min_samples_leaf=2, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=20, total=   0.9s\n",
      "[CV] min_samples_leaf=2, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=40, total=   1.7s\n",
      "[CV] min_samples_leaf=2, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=40, total=   1.7s\n",
      "[CV] min_samples_leaf=2, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=40, total=   1.4s\n",
      "[CV] min_samples_leaf=2, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=80, total=   3.5s\n",
      "[CV] min_samples_leaf=2, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=80, total=   3.3s\n",
      "[CV] min_samples_leaf=2, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=80, total=   3.0s\n",
      "[CV] min_samples_leaf=2, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=2, n_estimators=160, total=   5.8s\n",
      "[CV] min_samples_leaf=2, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=2, n_estimators=160, total=   6.4s\n",
      "[CV] min_samples_leaf=2, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=2, n_estimators=160, total=   6.1s\n",
      "[CV] min_samples_leaf=4, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=20, total=   0.5s\n",
      "[CV] min_samples_leaf=4, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=20, total=   0.5s\n",
      "[CV] min_samples_leaf=4, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=20, total=   0.5s\n",
      "[CV] min_samples_leaf=4, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=40, total=   1.2s\n",
      "[CV] min_samples_leaf=4, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=40, total=   1.3s\n",
      "[CV] min_samples_leaf=4, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=40, total=   1.3s\n",
      "[CV] min_samples_leaf=4, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=80, total=   2.3s\n",
      "[CV] min_samples_leaf=4, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=80, total=   2.4s\n",
      "[CV] min_samples_leaf=4, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=80, total=   2.4s\n",
      "[CV] min_samples_leaf=4, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=4, n_estimators=160, total=   4.8s\n",
      "[CV] min_samples_leaf=4, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=4, n_estimators=160, total=   5.0s\n",
      "[CV] min_samples_leaf=4, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=4, n_estimators=160, total=   5.0s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done  36 out of  36 | elapsed:  1.8min finished\n"
     ]
    }
   ],
   "source": [
    "rf = RandomForestRegressor(random_state=42)\n",
    "\n",
    "# Note: first batch of n_estimators started at 5 and returned optimal solution at the max value (40).\n",
    "# So I have increased this hyperparameter.\n",
    "\n",
    "parameters = {'n_estimators':[20,40,80,160],\n",
    "              #'max_depth':[4,8,16,32,64],\n",
    "              'min_samples_leaf':[1,2,4]\n",
    "             }\n",
    "\n",
    "# TODO: Make a scoring object using make_scorer()\n",
    "scorer = make_scorer(r2_score)\n",
    "\n",
    "# TODO: Perform grid search on the regressor using 'scorer' as the scoring method using GridSearchCV()\n",
    "grid_obj = GridSearchCV(rf, param_grid=parameters, scoring=scorer, verbose = 2)\n",
    "\n",
    "# TODO: Fit the grid search object to the training data and find the optimal parameters using fit()\n",
    "grid_fit = grid_obj.fit(X_train, y_train)\n",
    "\n",
    "# Get the estimator\n",
    "best_rf = grid_fit.best_estimator_\n",
    "\n",
    "# Make predictions using the unoptimized and model\n",
    "y_pred = best_rf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=None,\n",
       "           max_features='auto', max_leaf_nodes=None,\n",
       "           min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "           min_samples_leaf=2, min_samples_split=2,\n",
       "           min_weight_fraction_leaf=0.0, n_estimators=160, n_jobs=1,\n",
       "           oob_score=False, random_state=42, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_rf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Actuals')"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(y_pred, y_test)\n",
    "plt.xlabel('Predictions')\n",
    "plt.ylabel('Actuals')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "feature_plot_h(best_rf.feature_importances_, X_train, 5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Evaluate the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7450639301341384"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "r2_score(y_pred, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "An R2 of 0.745 is phenomenal! But a lot of importance is definitely coming from the first few features. \n",
    "\n",
    "## Feature refinement\n",
    "\n",
    "With so many features I'm a little concerned about overlap, so to check this let's cut the features used for modelling back to just the top 75 (remembering the original set had ~450)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>REGISTERED MOTOR VEHICLES - NUMBER BY TYPE OF VEHICLE Heavy rigid trucks no.</th>\n",
       "      <th>REGISTERED MOTOR VEHICLES - NUMBER BY TYPE OF VEHICLE Articulated trucks no.</th>\n",
       "      <th>EMPLOYED BY INDUSTRY Agriculture, Forestry and Fishing %</th>\n",
       "      <th>EMPLOYED BY INDUSTRY Manufacturing %</th>\n",
       "      <th>EMPLOYED BY INDUSTRY Electricity, Gas, Water &amp; Waste Services %</th>\n",
       "      <th>EMPLOYED BY INDUSTRY Wholesale Trade %</th>\n",
       "      <th>EMPLOYED BY INDUSTRY Retail Trade %</th>\n",
       "      <th>EMPLOYED BY INDUSTRY Financial &amp; Insurance Services %</th>\n",
       "      <th>EMPLOYED BY INDUSTRY Education &amp; Training %</th>\n",
       "      <th>PERSONS WITH POST SCHOOL QUALIFICATIONS Advanced Diploma, or Diploma  %</th>\n",
       "      <th>...</th>\n",
       "      <th>ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  Working Age Population (15-64 years) %</th>\n",
       "      <th>BIRTHS AND DEATHS Total Fertility Rate (per female)  rate</th>\n",
       "      <th>POPULATION DENSITY Population density  persons/km2</th>\n",
       "      <th>ABORIGINAL AND TORRES STRAIT ISLANDER PEOPLES Proportion of total population  %</th>\n",
       "      <th>OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Born in North-West Europe  %</th>\n",
       "      <th>OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Born in Southern &amp; Eastern Europe  %</th>\n",
       "      <th>OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Born in South-East Asia  %</th>\n",
       "      <th>OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Total Born Overseas  %</th>\n",
       "      <th>AUSTRALIAN CITIZENSHIP Not an Australian citizen  %</th>\n",
       "      <th>lat</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CODE</th>\n",
       "      <th>LABEL</th>\n",
       "      <th>YEAR</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>112011241</th>\n",
       "      <th>Evans Head</th>\n",
       "      <th>2016</th>\n",
       "      <td>83.0</td>\n",
       "      <td>46.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>7.9</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1.6</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1.3</td>\n",
       "      <td>10.2</td>\n",
       "      <td>6.9</td>\n",
       "      <td>...</td>\n",
       "      <td>58.7</td>\n",
       "      <td>1.9</td>\n",
       "      <td>9.8</td>\n",
       "      <td>4.5</td>\n",
       "      <td>3.3</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.5</td>\n",
       "      <td>6.6</td>\n",
       "      <td>2.8</td>\n",
       "      <td>-29.117409</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126021591</th>\n",
       "      <th>Ryde</th>\n",
       "      <th>2016</th>\n",
       "      <td>64.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.6</td>\n",
       "      <td>4.8</td>\n",
       "      <td>9.3</td>\n",
       "      <td>6.9</td>\n",
       "      <td>7.5</td>\n",
       "      <td>9.8</td>\n",
       "      <td>...</td>\n",
       "      <td>68.8</td>\n",
       "      <td>1.6</td>\n",
       "      <td>3894.9</td>\n",
       "      <td>0.4</td>\n",
       "      <td>3.3</td>\n",
       "      <td>5.0</td>\n",
       "      <td>6.2</td>\n",
       "      <td>45.4</td>\n",
       "      <td>20.1</td>\n",
       "      <td>-33.801575</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021011</th>\n",
       "      <th>Queanbeyan Region</th>\n",
       "      <th>2016</th>\n",
       "      <td>404.0</td>\n",
       "      <td>33.0</td>\n",
       "      <td>2.7</td>\n",
       "      <td>3.2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.5</td>\n",
       "      <td>6.2</td>\n",
       "      <td>1.3</td>\n",
       "      <td>7.9</td>\n",
       "      <td>10.9</td>\n",
       "      <td>...</td>\n",
       "      <td>67.5</td>\n",
       "      <td>1.9</td>\n",
       "      <td>5.7</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.3</td>\n",
       "      <td>1.5</td>\n",
       "      <td>0.8</td>\n",
       "      <td>13.0</td>\n",
       "      <td>2.6</td>\n",
       "      <td>-35.352466</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116031315</th>\n",
       "      <th>Hassall Grove - Plumpton</th>\n",
       "      <th>2016</th>\n",
       "      <td>197.0</td>\n",
       "      <td>44.0</td>\n",
       "      <td>0.4</td>\n",
       "      <td>10.9</td>\n",
       "      <td>0.9</td>\n",
       "      <td>3.9</td>\n",
       "      <td>11.2</td>\n",
       "      <td>6.0</td>\n",
       "      <td>5.3</td>\n",
       "      <td>8.9</td>\n",
       "      <td>...</td>\n",
       "      <td>69.9</td>\n",
       "      <td>1.8</td>\n",
       "      <td>3435.0</td>\n",
       "      <td>2.6</td>\n",
       "      <td>1.7</td>\n",
       "      <td>2.2</td>\n",
       "      <td>15.5</td>\n",
       "      <td>42.8</td>\n",
       "      <td>10.1</td>\n",
       "      <td>-33.731667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125021477</th>\n",
       "      <th>Ermington - Rydalmere</th>\n",
       "      <th>2016</th>\n",
       "      <td>97.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.1</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.8</td>\n",
       "      <td>4.7</td>\n",
       "      <td>9.1</td>\n",
       "      <td>5.4</td>\n",
       "      <td>7.7</td>\n",
       "      <td>9.8</td>\n",
       "      <td>...</td>\n",
       "      <td>67.2</td>\n",
       "      <td>1.9</td>\n",
       "      <td>2233.7</td>\n",
       "      <td>0.9</td>\n",
       "      <td>3.3</td>\n",
       "      <td>2.4</td>\n",
       "      <td>4.7</td>\n",
       "      <td>39.9</td>\n",
       "      <td>14.1</td>\n",
       "      <td>-33.817674</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 75 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         REGISTERED MOTOR VEHICLES - NUMBER BY TYPE OF VEHICLE Heavy rigid trucks no.   \\\n",
       "CODE      LABEL                    YEAR                                                                                  \n",
       "112011241 Evans Head               2016                                               83.0                               \n",
       "126021591 Ryde                     2016                                               64.0                               \n",
       "101021011 Queanbeyan Region        2016                                              404.0                               \n",
       "116031315 Hassall Grove - Plumpton 2016                                              197.0                               \n",
       "125021477 Ermington - Rydalmere    2016                                               97.0                               \n",
       "\n",
       "                                         REGISTERED MOTOR VEHICLES - NUMBER BY TYPE OF VEHICLE Articulated trucks no.   \\\n",
       "CODE      LABEL                    YEAR                                                                                  \n",
       "112011241 Evans Head               2016                                               46.0                               \n",
       "126021591 Ryde                     2016                                                2.0                               \n",
       "101021011 Queanbeyan Region        2016                                               33.0                               \n",
       "116031315 Hassall Grove - Plumpton 2016                                               44.0                               \n",
       "125021477 Ermington - Rydalmere    2016                                                5.0                               \n",
       "\n",
       "                                         EMPLOYED BY INDUSTRY Agriculture, Forestry and Fishing %  \\\n",
       "CODE      LABEL                    YEAR                                                             \n",
       "112011241 Evans Head               2016                                                6.0          \n",
       "126021591 Ryde                     2016                                                0.1          \n",
       "101021011 Queanbeyan Region        2016                                                2.7          \n",
       "116031315 Hassall Grove - Plumpton 2016                                                0.4          \n",
       "125021477 Ermington - Rydalmere    2016                                                0.1          \n",
       "\n",
       "                                         EMPLOYED BY INDUSTRY Manufacturing %  \\\n",
       "CODE      LABEL                    YEAR                                         \n",
       "112011241 Evans Head               2016                                   7.9   \n",
       "126021591 Ryde                     2016                                   5.0   \n",
       "101021011 Queanbeyan Region        2016                                   3.2   \n",
       "116031315 Hassall Grove - Plumpton 2016                                  10.9   \n",
       "125021477 Ermington - Rydalmere    2016                                   6.0   \n",
       "\n",
       "                                         EMPLOYED BY INDUSTRY Electricity, Gas, Water & Waste Services %  \\\n",
       "CODE      LABEL                    YEAR                                                                    \n",
       "112011241 Evans Head               2016                                                1.2                 \n",
       "126021591 Ryde                     2016                                                0.6                 \n",
       "101021011 Queanbeyan Region        2016                                                1.0                 \n",
       "116031315 Hassall Grove - Plumpton 2016                                                0.9                 \n",
       "125021477 Ermington - Rydalmere    2016                                                0.8                 \n",
       "\n",
       "                                         EMPLOYED BY INDUSTRY Wholesale Trade %  \\\n",
       "CODE      LABEL                    YEAR                                           \n",
       "112011241 Evans Head               2016                                     1.6   \n",
       "126021591 Ryde                     2016                                     4.8   \n",
       "101021011 Queanbeyan Region        2016                                     1.5   \n",
       "116031315 Hassall Grove - Plumpton 2016                                     3.9   \n",
       "125021477 Ermington - Rydalmere    2016                                     4.7   \n",
       "\n",
       "                                         EMPLOYED BY INDUSTRY Retail Trade %  \\\n",
       "CODE      LABEL                    YEAR                                        \n",
       "112011241 Evans Head               2016                                 10.0   \n",
       "126021591 Ryde                     2016                                  9.3   \n",
       "101021011 Queanbeyan Region        2016                                  6.2   \n",
       "116031315 Hassall Grove - Plumpton 2016                                 11.2   \n",
       "125021477 Ermington - Rydalmere    2016                                  9.1   \n",
       "\n",
       "                                         EMPLOYED BY INDUSTRY Financial & Insurance Services %  \\\n",
       "CODE      LABEL                    YEAR                                                          \n",
       "112011241 Evans Head               2016                                                1.3       \n",
       "126021591 Ryde                     2016                                                6.9       \n",
       "101021011 Queanbeyan Region        2016                                                1.3       \n",
       "116031315 Hassall Grove - Plumpton 2016                                                6.0       \n",
       "125021477 Ermington - Rydalmere    2016                                                5.4       \n",
       "\n",
       "                                         EMPLOYED BY INDUSTRY Education & Training %  \\\n",
       "CODE      LABEL                    YEAR                                                \n",
       "112011241 Evans Head               2016                                         10.2   \n",
       "126021591 Ryde                     2016                                          7.5   \n",
       "101021011 Queanbeyan Region        2016                                          7.9   \n",
       "116031315 Hassall Grove - Plumpton 2016                                          5.3   \n",
       "125021477 Ermington - Rydalmere    2016                                          7.7   \n",
       "\n",
       "                                         PERSONS WITH POST SCHOOL QUALIFICATIONS Advanced Diploma, or Diploma  %  \\\n",
       "CODE      LABEL                    YEAR                                                                            \n",
       "112011241 Evans Head               2016                                                6.9                         \n",
       "126021591 Ryde                     2016                                                9.8                         \n",
       "101021011 Queanbeyan Region        2016                                               10.9                         \n",
       "116031315 Hassall Grove - Plumpton 2016                                                8.9                         \n",
       "125021477 Ermington - Rydalmere    2016                                                9.8                         \n",
       "\n",
       "                                           ...      \\\n",
       "CODE      LABEL                    YEAR    ...       \n",
       "112011241 Evans Head               2016    ...       \n",
       "126021591 Ryde                     2016    ...       \n",
       "101021011 Queanbeyan Region        2016    ...       \n",
       "116031315 Hassall Grove - Plumpton 2016    ...       \n",
       "125021477 Ermington - Rydalmere    2016    ...       \n",
       "\n",
       "                                         ESTIMATED RESIDENT POPULATION (USUAL RESIDENTS, ERP AT 30 JUNE)  Working Age Population (15-64 years) %  \\\n",
       "CODE      LABEL                    YEAR                                                                                                            \n",
       "112011241 Evans Head               2016                                               58.7                                                         \n",
       "126021591 Ryde                     2016                                               68.8                                                         \n",
       "101021011 Queanbeyan Region        2016                                               67.5                                                         \n",
       "116031315 Hassall Grove - Plumpton 2016                                               69.9                                                         \n",
       "125021477 Ermington - Rydalmere    2016                                               67.2                                                         \n",
       "\n",
       "                                         BIRTHS AND DEATHS Total Fertility Rate (per female)  rate  \\\n",
       "CODE      LABEL                    YEAR                                                              \n",
       "112011241 Evans Head               2016                                                1.9           \n",
       "126021591 Ryde                     2016                                                1.6           \n",
       "101021011 Queanbeyan Region        2016                                                1.9           \n",
       "116031315 Hassall Grove - Plumpton 2016                                                1.8           \n",
       "125021477 Ermington - Rydalmere    2016                                                1.9           \n",
       "\n",
       "                                         POPULATION DENSITY Population density  persons/km2  \\\n",
       "CODE      LABEL                    YEAR                                                       \n",
       "112011241 Evans Head               2016                                                9.8    \n",
       "126021591 Ryde                     2016                                             3894.9    \n",
       "101021011 Queanbeyan Region        2016                                                5.7    \n",
       "116031315 Hassall Grove - Plumpton 2016                                             3435.0    \n",
       "125021477 Ermington - Rydalmere    2016                                             2233.7    \n",
       "\n",
       "                                         ABORIGINAL AND TORRES STRAIT ISLANDER PEOPLES Proportion of total population  %  \\\n",
       "CODE      LABEL                    YEAR                                                                                    \n",
       "112011241 Evans Head               2016                                                4.5                                 \n",
       "126021591 Ryde                     2016                                                0.4                                 \n",
       "101021011 Queanbeyan Region        2016                                                2.0                                 \n",
       "116031315 Hassall Grove - Plumpton 2016                                                2.6                                 \n",
       "125021477 Ermington - Rydalmere    2016                                                0.9                                 \n",
       "\n",
       "                                         OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Born in North-West Europe  %  \\\n",
       "CODE      LABEL                    YEAR                                                                                           \n",
       "112011241 Evans Head               2016                                                3.3                                        \n",
       "126021591 Ryde                     2016                                                3.3                                        \n",
       "101021011 Queanbeyan Region        2016                                                6.3                                        \n",
       "116031315 Hassall Grove - Plumpton 2016                                                1.7                                        \n",
       "125021477 Ermington - Rydalmere    2016                                                3.3                                        \n",
       "\n",
       "                                         OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Born in Southern & Eastern Europe  %  \\\n",
       "CODE      LABEL                    YEAR                                                                                                   \n",
       "112011241 Evans Head               2016                                                0.2                                                \n",
       "126021591 Ryde                     2016                                                5.0                                                \n",
       "101021011 Queanbeyan Region        2016                                                1.5                                                \n",
       "116031315 Hassall Grove - Plumpton 2016                                                2.2                                                \n",
       "125021477 Ermington - Rydalmere    2016                                                2.4                                                \n",
       "\n",
       "                                         OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Born in South-East Asia  %  \\\n",
       "CODE      LABEL                    YEAR                                                                                         \n",
       "112011241 Evans Head               2016                                                0.5                                      \n",
       "126021591 Ryde                     2016                                                6.2                                      \n",
       "101021011 Queanbeyan Region        2016                                                0.8                                      \n",
       "116031315 Hassall Grove - Plumpton 2016                                               15.5                                      \n",
       "125021477 Ermington - Rydalmere    2016                                                4.7                                      \n",
       "\n",
       "                                         OVERSEAS BORN POPULATION (PROPORTION OF TOTAL POPULATION) Total Born Overseas  %  \\\n",
       "CODE      LABEL                    YEAR                                                                                     \n",
       "112011241 Evans Head               2016                                                6.6                                  \n",
       "126021591 Ryde                     2016                                               45.4                                  \n",
       "101021011 Queanbeyan Region        2016                                               13.0                                  \n",
       "116031315 Hassall Grove - Plumpton 2016                                               42.8                                  \n",
       "125021477 Ermington - Rydalmere    2016                                               39.9                                  \n",
       "\n",
       "                                         AUSTRALIAN CITIZENSHIP Not an Australian citizen  %  \\\n",
       "CODE      LABEL                    YEAR                                                        \n",
       "112011241 Evans Head               2016                                                2.8     \n",
       "126021591 Ryde                     2016                                               20.1     \n",
       "101021011 Queanbeyan Region        2016                                                2.6     \n",
       "116031315 Hassall Grove - Plumpton 2016                                               10.1     \n",
       "125021477 Ermington - Rydalmere    2016                                               14.1     \n",
       "\n",
       "                                               lat  \n",
       "CODE      LABEL                    YEAR             \n",
       "112011241 Evans Head               2016 -29.117409  \n",
       "126021591 Ryde                     2016 -33.801575  \n",
       "101021011 Queanbeyan Region        2016 -35.352466  \n",
       "116031315 Hassall Grove - Plumpton 2016 -33.731667  \n",
       "125021477 Ermington - Rydalmere    2016 -33.817674  \n",
       "\n",
       "[5 rows x 75 columns]"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Keep only top 75 features\n",
    "indices = np.argsort(best_rf.feature_importances_)[::-1]\n",
    "columns = X_train.columns.values[indices[75:]]\n",
    "\n",
    "drop_cols = list(columns)\n",
    "X_train_Refined = X_train.copy()\n",
    "X_train_Refined.drop(drop_cols, axis=1, inplace=True)\n",
    "X_train_Refined.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 20 candidates, totalling 60 fits\n",
      "[CV] min_samples_leaf=1, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=20, total=   0.1s\n",
      "[CV] min_samples_leaf=1, n_estimators=20 .............................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.1s remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] .............. min_samples_leaf=1, n_estimators=20, total=   0.1s\n",
      "[CV] min_samples_leaf=1, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=20, total=   0.1s\n",
      "[CV] min_samples_leaf=1, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=1, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=1, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=1, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=80, total=   0.6s\n",
      "[CV] min_samples_leaf=1, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=80, total=   0.7s\n",
      "[CV] min_samples_leaf=1, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=1, n_estimators=80, total=   0.7s\n",
      "[CV] min_samples_leaf=1, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=1, n_estimators=160, total=   1.2s\n",
      "[CV] min_samples_leaf=1, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=1, n_estimators=160, total=   1.4s\n",
      "[CV] min_samples_leaf=1, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=1, n_estimators=160, total=   1.8s\n",
      "[CV] min_samples_leaf=2, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=20, total=   0.1s\n",
      "[CV] min_samples_leaf=2, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=20, total=   0.1s\n",
      "[CV] min_samples_leaf=2, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=20, total=   0.0s\n",
      "[CV] min_samples_leaf=2, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=2, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=2, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=2, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=80, total=   0.6s\n",
      "[CV] min_samples_leaf=2, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=80, total=   0.5s\n",
      "[CV] min_samples_leaf=2, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=2, n_estimators=80, total=   0.5s\n",
      "[CV] min_samples_leaf=2, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=2, n_estimators=160, total=   1.0s\n",
      "[CV] min_samples_leaf=2, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=2, n_estimators=160, total=   1.1s\n",
      "[CV] min_samples_leaf=2, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=2, n_estimators=160, total=   1.2s\n",
      "[CV] min_samples_leaf=3, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=3, n_estimators=20, total=   0.0s\n",
      "[CV] min_samples_leaf=3, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=3, n_estimators=20, total=   0.0s\n",
      "[CV] min_samples_leaf=3, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=3, n_estimators=20, total=   0.0s\n",
      "[CV] min_samples_leaf=3, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=3, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=3, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=3, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=3, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=3, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=3, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=3, n_estimators=80, total=   0.4s\n",
      "[CV] min_samples_leaf=3, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=3, n_estimators=80, total=   0.4s\n",
      "[CV] min_samples_leaf=3, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=3, n_estimators=80, total=   0.5s\n",
      "[CV] min_samples_leaf=3, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=3, n_estimators=160, total=   1.3s\n",
      "[CV] min_samples_leaf=3, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=3, n_estimators=160, total=   0.9s\n",
      "[CV] min_samples_leaf=3, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=3, n_estimators=160, total=   1.0s\n",
      "[CV] min_samples_leaf=4, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=20, total=   0.0s\n",
      "[CV] min_samples_leaf=4, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=20, total=   0.0s\n",
      "[CV] min_samples_leaf=4, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=20, total=   0.0s\n",
      "[CV] min_samples_leaf=4, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=40, total=   0.1s\n",
      "[CV] min_samples_leaf=4, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=4, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=4, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=80, total=   0.4s\n",
      "[CV] min_samples_leaf=4, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=80, total=   0.4s\n",
      "[CV] min_samples_leaf=4, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=4, n_estimators=80, total=   0.4s\n",
      "[CV] min_samples_leaf=4, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=4, n_estimators=160, total=   1.1s\n",
      "[CV] min_samples_leaf=4, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=4, n_estimators=160, total=   0.9s\n",
      "[CV] min_samples_leaf=4, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=4, n_estimators=160, total=   0.9s\n",
      "[CV] min_samples_leaf=5, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=5, n_estimators=20, total=   0.0s\n",
      "[CV] min_samples_leaf=5, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=5, n_estimators=20, total=   0.1s\n",
      "[CV] min_samples_leaf=5, n_estimators=20 .............................\n",
      "[CV] .............. min_samples_leaf=5, n_estimators=20, total=   0.0s\n",
      "[CV] min_samples_leaf=5, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=5, n_estimators=40, total=   0.1s\n",
      "[CV] min_samples_leaf=5, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=5, n_estimators=40, total=   0.1s\n",
      "[CV] min_samples_leaf=5, n_estimators=40 .............................\n",
      "[CV] .............. min_samples_leaf=5, n_estimators=40, total=   0.2s\n",
      "[CV] min_samples_leaf=5, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=5, n_estimators=80, total=   0.4s\n",
      "[CV] min_samples_leaf=5, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=5, n_estimators=80, total=   0.4s\n",
      "[CV] min_samples_leaf=5, n_estimators=80 .............................\n",
      "[CV] .............. min_samples_leaf=5, n_estimators=80, total=   0.4s\n",
      "[CV] min_samples_leaf=5, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=5, n_estimators=160, total=   0.9s\n",
      "[CV] min_samples_leaf=5, n_estimators=160 ............................\n",
      "[CV] ............. min_samples_leaf=5, n_estimators=160, total=   0.8s\n",
      "[CV] min_samples_leaf=5, n_estimators=160 ............................\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ............. min_samples_leaf=5, n_estimators=160, total=   0.8s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done  60 out of  60 | elapsed:   33.5s finished\n"
     ]
    }
   ],
   "source": [
    "rf = RandomForestRegressor(random_state=42)\n",
    "\n",
    "parameters = {'n_estimators':[20,40,80,160],\n",
    "              #'max_depth':[4,8,16,32,64],\n",
    "              'min_samples_leaf':[1,2,3,4,5]\n",
    "             }\n",
    "\n",
    "# TODO: Make a scoring object using make_scorer()\n",
    "scorer = make_scorer(r2_score)\n",
    "\n",
    "# TODO: Perform grid search on the regressor using 'scorer' as the scoring method using GridSearchCV()\n",
    "grid_obj = GridSearchCV(rf, param_grid=parameters, scoring=scorer, verbose = 2)\n",
    "\n",
    "# TODO: Fit the grid search object to the training data and find the optimal parameters using fit()\n",
    "grid_fit = grid_obj.fit(X_train_Refined, y_train)\n",
    "\n",
    "# Get the estimator\n",
    "best_rf = grid_fit.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make predictions using the unoptimized and model\n",
    "y_pred = best_rf.predict(X_test.drop(drop_cols, axis=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=None,\n",
       "           max_features='auto', max_leaf_nodes=None,\n",
       "           min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "           min_samples_leaf=2, min_samples_split=2,\n",
       "           min_weight_fraction_leaf=0.0, n_estimators=160, n_jobs=1,\n",
       "           oob_score=False, random_state=42, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "best_rf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Actuals')"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(y_pred, y_test)\n",
    "plt.xlabel('Predictions')\n",
    "plt.ylabel('Actuals')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7439723980779221"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "r2_score(y_pred, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "feature_plot_h(best_rf.feature_importances_, X_train.drop(drop_cols, axis=1), 10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Just... wow! The R2 barely dropped at all, and our key features barely changed.\n",
    "\n",
    "## Feature impact analysis - \"Deploy\"\n",
    "\n",
    "But what do these features actually mean? If I increase population density does it decrease or increase solar panel installations (in line with the correlations above)? Or is it multi dimensional where it decreases then increases? By changing the values in the training set then re-predicting, we can model this out quite nicely.\n",
    " \n",
    "I'm choosing to change the base values over a range of +/- the standard deviation of the original feature.\n",
    "\n",
    "Note - I am choosing to reimport everything at this point for ease of jumping in and out of the workbook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Create X and y based on preprocessed data\n",
    "X = pd.read_csv('{}\\medians by code.csv'.format(nb_path), na_values='-', thousands=',')\n",
    "X.rename(columns={'lat':'LATITUDE'}, inplace=True)\n",
    "X.set_index(['CODE', 'LABEL', 'YEAR'], inplace=True)\n",
    "y = X['Solar Per Household'].copy()\n",
    "X.drop('Solar Per Household', axis=1, inplace=True)\n",
    "\n",
    "# Split the 'features' and 'income' data into training and testing sets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7450028827166401"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Initialise RF based on GridSearchCV parameters identified above\n",
    "rf = RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=None,\n",
    "           max_features='auto', max_leaf_nodes=None,\n",
    "           min_impurity_decrease=0.0, min_impurity_split=None,\n",
    "           min_samples_leaf=2, min_samples_split=2,\n",
    "           min_weight_fraction_leaf=0.0, n_estimators=160, n_jobs=None,\n",
    "           oob_score=False, random_state=42, verbose=0, warm_start=False)\n",
    "\n",
    "rf.fit(X_train, y_train)\n",
    "\n",
    "# Make predictions using the model\n",
    "y_pred = rf.predict(X_test)\n",
    "r2_score(y_pred, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Keep only top 75 features\n",
    "indices = np.argsort(rf.feature_importances_)[::-1]\n",
    "worst_columns = list(X_train.columns.values[indices[75:]])\n",
    "\n",
    "X_train_Refined = X_train.copy()\n",
    "X_train_Refined.drop(worst_columns, axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7441132579464321"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Initialise another RF based on GridSearchCV parameters identified above for the leaner dataset\n",
    "lean_rf = RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=None,\n",
    "           max_features='auto', max_leaf_nodes=None,\n",
    "           min_impurity_decrease=0.0, min_impurity_split=None,\n",
    "           min_samples_leaf=2, min_samples_split=2,\n",
    "           min_weight_fraction_leaf=0.0, n_estimators=160, n_jobs=None,\n",
    "           oob_score=False, random_state=42, verbose=0, warm_start=False)\n",
    "\n",
    "lean_rf.fit(X_train_Refined, y_train)\n",
    "\n",
    "# Make predictions using the model\n",
    "y_pred = lean_rf.predict(X_test.drop(worst_columns, axis=1))\n",
    "r2_score(y_pred, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# What are the main indicators of a region having an increased propensity to have solar panels installed?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "feature_plot_h(lean_rf.feature_importances_, X_train_Refined, 5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# For these identified features, how do changes in the feature impact changes in solar panels installed?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "feature_impact_plot(lean_rf, X_train_Refined, 5, 'Solar per household')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# How have solar installations changed over time when using these features as categories?\n",
    "\n",
    "First we'll have to reload the source data and get insight into which of these features actually *have* multi-year data. So re-run the first 7 cells in the notebook (no cleaning beyond this!).\n",
    "\n",
    "After that I will break the features into quartiles and aggregate the data based on these features. Then plot the trend in gross installations over the timeframe available."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['EMPLOYED BY INDUSTRY Agriculture, Forestry and Fishing %',\n",
       " 'POPULATION DENSITY Population density  persons/km2',\n",
       " 'DWELLING STRUCTURE Flat or apartment  no.',\n",
       " 'LATITUDE',\n",
       " 'COMMUTING TO WORK Average commuting distance from Place of Usual Residence kms']"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indices = np.argsort(lean_rf.feature_importances_)[::-1]\n",
    "list(X_train_Refined.columns.values[indices[:5]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "top_features = list(X_train_Refined.columns.values[indices[:5]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2015    579\n",
       "2011    579\n",
       "2018    579\n",
       "2014    579\n",
       "2017    579\n",
       "2013    579\n",
       "2016    579\n",
       "2012    579\n",
       "Name: YEAR, dtype: int64"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load all the files into a single df\n",
    "df = pd.read_csv('{}\\CSV\\{}'.format(nb_path, files[0]), na_values='-', thousands=',')\n",
    "\n",
    "for file in range(1,len(files)):\n",
    "    df_temp = pd.read_csv('{}\\CSV\\{}'.format(nb_path, files[file]), na_values='-', thousands=',')\n",
    "    df = pd.merge(df, df_temp, how='outer', left_on=['CODE','YEAR','LABEL'], right_on=['CODE','YEAR','LABEL'])\n",
    "\n",
    "# Load in latitude and longitude data\n",
    "df_temp = pd.read_csv('{}\\latlng.csv'.format(nb_path), na_values='-', thousands=',')\n",
    "df_temp.rename(columns = {df_temp.columns[0] : 'LABEL'}, inplace=True)\n",
    "df = pd.merge(df, df_temp, how='inner', left_on='LABEL', right_on='LABEL')\n",
    "df.rename(columns={'lat':'LATITUDE'}, inplace=True)\n",
    "\n",
    "df['YEAR'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>CODE</th>\n",
       "      <th>LABEL</th>\n",
       "      <th>YEAR</th>\n",
       "      <th>NUMBER OF BUSINESSES Number of non-employing businesses no.</th>\n",
       "      <th>NUMBER OF BUSINESSES 1-4 employees no.</th>\n",
       "      <th>NUMBER OF BUSINESSES 5-19 employees no.</th>\n",
       "      <th>NUMBER OF BUSINESSES 20 or more employees no.</th>\n",
       "      <th>NUMBER OF BUSINESSES Total number of businesses no.</th>\n",
       "      <th>BUSINESS ENTRIES Number of non-employing business entries  no.</th>\n",
       "      <th>BUSINESS ENTRIES 1-4 employees no.</th>\n",
       "      <th>...</th>\n",
       "      <th>RELIGIOUS AFFILIATION Judaism   %</th>\n",
       "      <th>RELIGIOUS AFFILIATION Other Religions   %</th>\n",
       "      <th>RELIGIOUS AFFILIATION Secular Beliefs, Other Spiritual Beliefs and No Religious Affiliation  %</th>\n",
       "      <th>RELIGIOUS AFFILIATION Inadequately described or not stated   %</th>\n",
       "      <th>AUSTRALIAN CITIZENSHIP Australian citizen  %</th>\n",
       "      <th>AUSTRALIAN CITIZENSHIP Not an Australian citizen  %</th>\n",
       "      <th>AUSTRALIAN CITIZENSHIP Australian Citizenship - Not stated  %</th>\n",
       "      <th>SPEAKS A LANGUAGE OTHER THAN ENGLISH AT HOME Proportion of total population  %</th>\n",
       "      <th>LATITUDE</th>\n",
       "      <th>long</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>101021009</td>\n",
       "      <td>Queanbeyan</td>\n",
       "      <td>2011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>84.9</td>\n",
       "      <td>8.2</td>\n",
       "      <td>6.8</td>\n",
       "      <td>18.6</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>149.233725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>101021009</td>\n",
       "      <td>Queanbeyan</td>\n",
       "      <td>2012</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>149.233725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>101021009</td>\n",
       "      <td>Queanbeyan</td>\n",
       "      <td>2013</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>149.233725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>101021009</td>\n",
       "      <td>Queanbeyan</td>\n",
       "      <td>2014</td>\n",
       "      <td>398.0</td>\n",
       "      <td>210.0</td>\n",
       "      <td>84.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>715.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>149.233725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>101021009</td>\n",
       "      <td>Queanbeyan</td>\n",
       "      <td>2015</td>\n",
       "      <td>401.0</td>\n",
       "      <td>185.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>691.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>149.233725</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 589 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         CODE       LABEL  YEAR  \\\n",
       "8   101021009  Queanbeyan  2011   \n",
       "9   101021009  Queanbeyan  2012   \n",
       "10  101021009  Queanbeyan  2013   \n",
       "11  101021009  Queanbeyan  2014   \n",
       "12  101021009  Queanbeyan  2015   \n",
       "\n",
       "    NUMBER OF BUSINESSES Number of non-employing businesses no.  \\\n",
       "8                                                 NaN             \n",
       "9                                                 NaN             \n",
       "10                                                NaN             \n",
       "11                                              398.0             \n",
       "12                                              401.0             \n",
       "\n",
       "    NUMBER OF BUSINESSES 1-4 employees no.  \\\n",
       "8                                      NaN   \n",
       "9                                      NaN   \n",
       "10                                     NaN   \n",
       "11                                   210.0   \n",
       "12                                   185.0   \n",
       "\n",
       "    NUMBER OF BUSINESSES 5-19 employees no.  \\\n",
       "8                                       NaN   \n",
       "9                                       NaN   \n",
       "10                                      NaN   \n",
       "11                                     84.0   \n",
       "12                                     82.0   \n",
       "\n",
       "    NUMBER OF BUSINESSES 20 or more employees no.  \\\n",
       "8                                             NaN   \n",
       "9                                             NaN   \n",
       "10                                            NaN   \n",
       "11                                           24.0   \n",
       "12                                           26.0   \n",
       "\n",
       "    NUMBER OF BUSINESSES Total number of businesses no.  \\\n",
       "8                                                 NaN     \n",
       "9                                                 NaN     \n",
       "10                                                NaN     \n",
       "11                                              715.0     \n",
       "12                                              691.0     \n",
       "\n",
       "    BUSINESS ENTRIES Number of non-employing business entries  no.  \\\n",
       "8                                                 NaN                \n",
       "9                                                 NaN                \n",
       "10                                                NaN                \n",
       "11                                                NaN                \n",
       "12                                               68.0                \n",
       "\n",
       "    BUSINESS ENTRIES 1-4 employees no.  ...  \\\n",
       "8                                  NaN  ...   \n",
       "9                                  NaN  ...   \n",
       "10                                 NaN  ...   \n",
       "11                                 NaN  ...   \n",
       "12                                19.0  ...   \n",
       "\n",
       "    RELIGIOUS AFFILIATION Judaism   %  \\\n",
       "8                                 NaN   \n",
       "9                                 NaN   \n",
       "10                                NaN   \n",
       "11                                NaN   \n",
       "12                                NaN   \n",
       "\n",
       "    RELIGIOUS AFFILIATION Other Religions   %  \\\n",
       "8                                         NaN   \n",
       "9                                         NaN   \n",
       "10                                        NaN   \n",
       "11                                        NaN   \n",
       "12                                        NaN   \n",
       "\n",
       "    RELIGIOUS AFFILIATION Secular Beliefs, Other Spiritual Beliefs and No Religious Affiliation  %  \\\n",
       "8                                                 NaN                                                \n",
       "9                                                 NaN                                                \n",
       "10                                                NaN                                                \n",
       "11                                                NaN                                                \n",
       "12                                                NaN                                                \n",
       "\n",
       "    RELIGIOUS AFFILIATION Inadequately described or not stated   %  \\\n",
       "8                                                 NaN                \n",
       "9                                                 NaN                \n",
       "10                                                NaN                \n",
       "11                                                NaN                \n",
       "12                                                NaN                \n",
       "\n",
       "    AUSTRALIAN CITIZENSHIP Australian citizen  %  \\\n",
       "8                                           84.9   \n",
       "9                                            NaN   \n",
       "10                                           NaN   \n",
       "11                                           NaN   \n",
       "12                                           NaN   \n",
       "\n",
       "    AUSTRALIAN CITIZENSHIP Not an Australian citizen  %  \\\n",
       "8                                                 8.2     \n",
       "9                                                 NaN     \n",
       "10                                                NaN     \n",
       "11                                                NaN     \n",
       "12                                                NaN     \n",
       "\n",
       "    AUSTRALIAN CITIZENSHIP Australian Citizenship - Not stated  %  \\\n",
       "8                                                 6.8               \n",
       "9                                                 NaN               \n",
       "10                                                NaN               \n",
       "11                                                NaN               \n",
       "12                                                NaN               \n",
       "\n",
       "    SPEAKS A LANGUAGE OTHER THAN ENGLISH AT HOME Proportion of total population  %  \\\n",
       "8                                                18.6                                \n",
       "9                                                 NaN                                \n",
       "10                                                NaN                                \n",
       "11                                                NaN                                \n",
       "12                                                NaN                                \n",
       "\n",
       "     LATITUDE        long  \n",
       "8  -35.352466  149.233725  \n",
       "9  -35.352466  149.233725  \n",
       "10 -35.352466  149.233725  \n",
       "11 -35.352466  149.233725  \n",
       "12 -35.352466  149.233725  \n",
       "\n",
       "[5 rows x 589 columns]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = remove_aggregation(df)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>YEAR</th>\n",
       "      <th>SOLAR INSTALLATIONS Small-Scale Solar Panel System Installations no.</th>\n",
       "      <th>EMPLOYED BY INDUSTRY Agriculture, Forestry and Fishing %</th>\n",
       "      <th>POPULATION DENSITY Population density  persons/km2</th>\n",
       "      <th>DWELLING STRUCTURE Flat or apartment  no.</th>\n",
       "      <th>LATITUDE</th>\n",
       "      <th>COMMUTING TO WORK Average commuting distance from Place of Usual Residence kms</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1452.0</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2012</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2282.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2013</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2319.8</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>2014</td>\n",
       "      <td>425.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2351.1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>2015</td>\n",
       "      <td>485.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2354.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>2016</td>\n",
       "      <td>548.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>2357.6</td>\n",
       "      <td>1564.0</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>10.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>2017</td>\n",
       "      <td>635.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2383.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>2018</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2405.4</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-35.352466</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>2011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>276.0</td>\n",
       "      <td>-33.285641</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>2012</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>566.5</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-33.285641</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    YEAR  \\\n",
       "8   2011   \n",
       "9   2012   \n",
       "10  2013   \n",
       "11  2014   \n",
       "12  2015   \n",
       "13  2016   \n",
       "14  2017   \n",
       "15  2018   \n",
       "24  2011   \n",
       "25  2012   \n",
       "\n",
       "    SOLAR INSTALLATIONS Small-Scale Solar Panel System Installations no.  \\\n",
       "8                                                 NaN                      \n",
       "9                                                 NaN                      \n",
       "10                                                NaN                      \n",
       "11                                              425.0                      \n",
       "12                                              485.0                      \n",
       "13                                              548.0                      \n",
       "14                                              635.0                      \n",
       "15                                                NaN                      \n",
       "24                                                NaN                      \n",
       "25                                                NaN                      \n",
       "\n",
       "    EMPLOYED BY INDUSTRY Agriculture, Forestry and Fishing %  \\\n",
       "8                                                 0.2          \n",
       "9                                                 NaN          \n",
       "10                                                NaN          \n",
       "11                                                NaN          \n",
       "12                                                NaN          \n",
       "13                                                0.2          \n",
       "14                                                NaN          \n",
       "15                                                NaN          \n",
       "24                                                1.0          \n",
       "25                                                NaN          \n",
       "\n",
       "    POPULATION DENSITY Population density  persons/km2  \\\n",
       "8                                                 NaN    \n",
       "9                                              2282.0    \n",
       "10                                             2319.8    \n",
       "11                                             2351.1    \n",
       "12                                             2354.0    \n",
       "13                                             2357.6    \n",
       "14                                             2383.0    \n",
       "15                                             2405.4    \n",
       "24                                                NaN    \n",
       "25                                              566.5    \n",
       "\n",
       "    DWELLING STRUCTURE Flat or apartment  no.   LATITUDE  \\\n",
       "8                                      1452.0 -35.352466   \n",
       "9                                         NaN -35.352466   \n",
       "10                                        NaN -35.352466   \n",
       "11                                        NaN -35.352466   \n",
       "12                                        NaN -35.352466   \n",
       "13                                     1564.0 -35.352466   \n",
       "14                                        NaN -35.352466   \n",
       "15                                        NaN -35.352466   \n",
       "24                                      276.0 -33.285641   \n",
       "25                                        NaN -33.285641   \n",
       "\n",
       "    COMMUTING TO WORK Average commuting distance from Place of Usual Residence kms  \n",
       "8                                                 NaN                               \n",
       "9                                                 NaN                               \n",
       "10                                                NaN                               \n",
       "11                                                NaN                               \n",
       "12                                                NaN                               \n",
       "13                                               10.7                               \n",
       "14                                                NaN                               \n",
       "15                                                NaN                               \n",
       "24                                                NaN                               \n",
       "25                                                NaN                               "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.loc[:,['YEAR', solar]+top_features]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So it looks like the only feature with data worth plotting as a key trend is 'Population density' over the course of 2014-2017, which to be honest is probably a great representative feature of the rest of the items anyway.   \n",
    "\n",
    "First, plot histograms for these 4 years for the features, just to check quartile movement isn't too drastic."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>YEAR</th>\n",
       "      <th>2014</th>\n",
       "      <th>2015</th>\n",
       "      <th>2016</th>\n",
       "      <th>2017</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CODE</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>101021007</th>\n",
       "      <td>343.0</td>\n",
       "      <td>380.0</td>\n",
       "      <td>418.0</td>\n",
       "      <td>466.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021008</th>\n",
       "      <td>357.0</td>\n",
       "      <td>407.0</td>\n",
       "      <td>460.0</td>\n",
       "      <td>534.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021009</th>\n",
       "      <td>425.0</td>\n",
       "      <td>485.0</td>\n",
       "      <td>548.0</td>\n",
       "      <td>635.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021010</th>\n",
       "      <td>193.0</td>\n",
       "      <td>220.0</td>\n",
       "      <td>249.0</td>\n",
       "      <td>288.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021011</th>\n",
       "      <td>799.0</td>\n",
       "      <td>923.0</td>\n",
       "      <td>1027.0</td>\n",
       "      <td>1189.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "YEAR        2014   2015    2016    2017\n",
       "CODE                                   \n",
       "101021007  343.0  380.0   418.0   466.0\n",
       "101021008  357.0  407.0   460.0   534.0\n",
       "101021009  425.0  485.0   548.0   635.0\n",
       "101021010  193.0  220.0   249.0   288.0\n",
       "101021011  799.0  923.0  1027.0  1189.0"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.pivot_table(df.loc[:,['CODE','YEAR', solar]], values = solar, columns = 'YEAR', index = 'CODE').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "histo_plots(pd.pivot_table(df.loc[:,['CODE','YEAR', solar]], values = solar, columns = 'YEAR', index = 'CODE'),4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "metric = 'POPULATION DENSITY Population density  persons/km2'\n",
    "pvt = pd.pivot_table(df.loc[:,['CODE','YEAR', metric]],\n",
    "                     values = metric, columns = 'YEAR', index = 'CODE').drop([2012,2013,2018],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "histo_plots(pvt, 4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So while the distribution for solar panels changed significantly over time, the population density itself does not appear to have moved much. Which indicates first of all that population density is unlikely to be a causative factor in the purchase decision of solar panels over time, but it is a large indicator of the *propensity* to purchase. To further explore this, let's plot the trend after segregating the data into quartiles."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>YEAR</th>\n",
       "      <th>2014</th>\n",
       "      <th>2015</th>\n",
       "      <th>2016</th>\n",
       "      <th>2017</th>\n",
       "      <th>2014 Quartile</th>\n",
       "      <th>2015 Quartile</th>\n",
       "      <th>2016 Quartile</th>\n",
       "      <th>2017 Quartile</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CODE</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>101021007</th>\n",
       "      <td>1.1</td>\n",
       "      <td>1.1</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1.2</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021008</th>\n",
       "      <td>1250.4</td>\n",
       "      <td>1232.1</td>\n",
       "      <td>1221.8</td>\n",
       "      <td>1215.0</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021009</th>\n",
       "      <td>2351.1</td>\n",
       "      <td>2354.0</td>\n",
       "      <td>2357.6</td>\n",
       "      <td>2383.0</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021010</th>\n",
       "      <td>382.0</td>\n",
       "      <td>381.5</td>\n",
       "      <td>382.2</td>\n",
       "      <td>385.7</td>\n",
       "      <td>q2</td>\n",
       "      <td>q2</td>\n",
       "      <td>q2</td>\n",
       "      <td>q2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021011</th>\n",
       "      <td>5.2</td>\n",
       "      <td>5.4</td>\n",
       "      <td>5.7</td>\n",
       "      <td>6.0</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "YEAR         2014    2015    2016    2017 2014 Quartile 2015 Quartile  \\\n",
       "CODE                                                                    \n",
       "101021007     1.1     1.1     1.2     1.2            q1            q1   \n",
       "101021008  1250.4  1232.1  1221.8  1215.0            q3            q3   \n",
       "101021009  2351.1  2354.0  2357.6  2383.0            q3            q3   \n",
       "101021010   382.0   381.5   382.2   385.7            q2            q2   \n",
       "101021011     5.2     5.4     5.7     6.0            q1            q1   \n",
       "\n",
       "YEAR      2016 Quartile 2017 Quartile  \n",
       "CODE                                   \n",
       "101021007            q1            q1  \n",
       "101021008            q3            q3  \n",
       "101021009            q3            q3  \n",
       "101021010            q2            q2  \n",
       "101021011            q1            q1  "
      ]
     },
     "execution_count": 115,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for col in pvt.columns:\n",
    "    pvt['{} Quartile'.format(col)] = pd.qcut(pvt[col], 4, labels = ['q1','q2','q3','q4'])\n",
    "\n",
    "pvt.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>YEAR</th>\n",
       "      <th>2014_x</th>\n",
       "      <th>2015_x</th>\n",
       "      <th>2016_x</th>\n",
       "      <th>2017_x</th>\n",
       "      <th>2014 Quartile</th>\n",
       "      <th>2015 Quartile</th>\n",
       "      <th>2016 Quartile</th>\n",
       "      <th>2017 Quartile</th>\n",
       "      <th>2014_y</th>\n",
       "      <th>2015_y</th>\n",
       "      <th>2016_y</th>\n",
       "      <th>2017_y</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CODE</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>101021007</th>\n",
       "      <td>1.1</td>\n",
       "      <td>1.1</td>\n",
       "      <td>1.2</td>\n",
       "      <td>1.2</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>343.0</td>\n",
       "      <td>380.0</td>\n",
       "      <td>418.0</td>\n",
       "      <td>466.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021008</th>\n",
       "      <td>1250.4</td>\n",
       "      <td>1232.1</td>\n",
       "      <td>1221.8</td>\n",
       "      <td>1215.0</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>357.0</td>\n",
       "      <td>407.0</td>\n",
       "      <td>460.0</td>\n",
       "      <td>534.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021009</th>\n",
       "      <td>2351.1</td>\n",
       "      <td>2354.0</td>\n",
       "      <td>2357.6</td>\n",
       "      <td>2383.0</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>q3</td>\n",
       "      <td>425.0</td>\n",
       "      <td>485.0</td>\n",
       "      <td>548.0</td>\n",
       "      <td>635.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021010</th>\n",
       "      <td>382.0</td>\n",
       "      <td>381.5</td>\n",
       "      <td>382.2</td>\n",
       "      <td>385.7</td>\n",
       "      <td>q2</td>\n",
       "      <td>q2</td>\n",
       "      <td>q2</td>\n",
       "      <td>q2</td>\n",
       "      <td>193.0</td>\n",
       "      <td>220.0</td>\n",
       "      <td>249.0</td>\n",
       "      <td>288.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101021011</th>\n",
       "      <td>5.2</td>\n",
       "      <td>5.4</td>\n",
       "      <td>5.7</td>\n",
       "      <td>6.0</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>q1</td>\n",
       "      <td>799.0</td>\n",
       "      <td>923.0</td>\n",
       "      <td>1027.0</td>\n",
       "      <td>1189.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "YEAR       2014_x  2015_x  2016_x  2017_x 2014 Quartile 2015 Quartile  \\\n",
       "CODE                                                                    \n",
       "101021007     1.1     1.1     1.2     1.2            q1            q1   \n",
       "101021008  1250.4  1232.1  1221.8  1215.0            q3            q3   \n",
       "101021009  2351.1  2354.0  2357.6  2383.0            q3            q3   \n",
       "101021010   382.0   381.5   382.2   385.7            q2            q2   \n",
       "101021011     5.2     5.4     5.7     6.0            q1            q1   \n",
       "\n",
       "YEAR      2016 Quartile 2017 Quartile  2014_y  2015_y  2016_y  2017_y  \n",
       "CODE                                                                   \n",
       "101021007            q1            q1   343.0   380.0   418.0   466.0  \n",
       "101021008            q3            q3   357.0   407.0   460.0   534.0  \n",
       "101021009            q3            q3   425.0   485.0   548.0   635.0  \n",
       "101021010            q2            q2   193.0   220.0   249.0   288.0  \n",
       "101021011            q1            q1   799.0   923.0  1027.0  1189.0  "
      ]
     },
     "execution_count": 117,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_temp = pd.pivot_table(df.loc[:,['CODE','YEAR', solar]],values = solar, columns = 'YEAR', index = 'CODE')\n",
    "pvt = pd.merge(pvt, df_temp, how='inner', left_on='CODE', right_on='CODE')\n",
    "pvt.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 67191., 100468.,  75456.,  49956.],\n",
       "       [ 77318., 111370.,  83925.,  53871.],\n",
       "       [ 85728., 121468.,  89646.,  59114.],\n",
       "       [ 97197., 137578.,  99123.,  65131.]])"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "year_list = ['2014_y','2015_y','2016_y','2017_y']\n",
    "q_list = ['q1','q2','q3','q4']\n",
    "pvt[pvt['2014 Quartile']=='q1']['2014_y'].sum()\n",
    "\n",
    "q_summary = np.zeros((4,4))\n",
    "\n",
    "for year in year_list:\n",
    "    for q in q_list:\n",
    "        q_summary[year_list.index(year), q_list.index(q)] = pvt[pvt[year.replace('_y',' Quartile')]==q][year].sum()\n",
    "\n",
    "q_summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>q1</th>\n",
       "      <th>q2</th>\n",
       "      <th>q3</th>\n",
       "      <th>q4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2014</th>\n",
       "      <td>67191.0</td>\n",
       "      <td>100468.0</td>\n",
       "      <td>75456.0</td>\n",
       "      <td>49956.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015</th>\n",
       "      <td>77318.0</td>\n",
       "      <td>111370.0</td>\n",
       "      <td>83925.0</td>\n",
       "      <td>53871.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016</th>\n",
       "      <td>85728.0</td>\n",
       "      <td>121468.0</td>\n",
       "      <td>89646.0</td>\n",
       "      <td>59114.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017</th>\n",
       "      <td>97197.0</td>\n",
       "      <td>137578.0</td>\n",
       "      <td>99123.0</td>\n",
       "      <td>65131.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           q1        q2       q3       q4\n",
       "2014  67191.0  100468.0  75456.0  49956.0\n",
       "2015  77318.0  111370.0  83925.0  53871.0\n",
       "2016  85728.0  121468.0  89646.0  59114.0\n",
       "2017  97197.0  137578.0  99123.0  65131.0"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_qsum = pd.DataFrame(q_summary, index = [y.replace('_y','') for y in year_list], columns = q_list)\n",
    "df_qsum"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>q1</th>\n",
       "      <th>q2</th>\n",
       "      <th>q3</th>\n",
       "      <th>q4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2014</th>\n",
       "      <td>22.926526</td>\n",
       "      <td>34.281113</td>\n",
       "      <td>25.746662</td>\n",
       "      <td>17.045699</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015</th>\n",
       "      <td>23.682018</td>\n",
       "      <td>34.111932</td>\n",
       "      <td>25.705701</td>\n",
       "      <td>16.500349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016</th>\n",
       "      <td>24.083876</td>\n",
       "      <td>34.124442</td>\n",
       "      <td>25.184573</td>\n",
       "      <td>16.607109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017</th>\n",
       "      <td>24.358380</td>\n",
       "      <td>34.478196</td>\n",
       "      <td>24.841052</td>\n",
       "      <td>16.322373</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             q1         q2         q3         q4\n",
       "2014  22.926526  34.281113  25.746662  17.045699\n",
       "2015  23.682018  34.111932  25.705701  16.500349\n",
       "2016  24.083876  34.124442  25.184573  16.607109\n",
       "2017  24.358380  34.478196  24.841052  16.322373"
      ]
     },
     "execution_count": 211,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_qsum.div(df_qsum.sum(axis=1),axis=0)*100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x1608018edd8>"
      ]
     },
     "execution_count": 213,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "trend_plot = df_qsum.plot(figsize=(10,7))\n",
    "\n",
    "trend_plot.set_title('Total solar installations \\n'\n",
    "                    + 'by population density quartile')\n",
    "trend_plot.tick_params(\n",
    "    axis='x',          # changes apply to the x-axis\n",
    "    which='both',      # both major and minor ticks are affected\n",
    "    bottom=False,      # ticks along the bottom edge are off\n",
    "    top=False,         # ticks along the top edge are off\n",
    "    labelbottom=True) # labels along the bottom edge are on\n",
    "trend_plot.set_xlabel('Year')\n",
    "trend_plot.set_ylabel('Number of solar panel installations')\n",
    "plt.rcParams.update({'font.size': 14})\n",
    "\n",
    "# Rename the legend features to make it a little less technical and more relatable to a general audience\n",
    "trend_plot.legend(labels=['Rural','Regional','Suburban','Urban'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "q1    0.446578\n",
       "q2    0.369371\n",
       "q3    0.313653\n",
       "q4    0.303767\n",
       "dtype: float64"
      ]
     },
     "execution_count": 148,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculate the total growth for each of the quartiles over the period 2014-2017\n",
    "df_qsum.iloc[3,:]/df_qsum.iloc[0,:] - 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "q1    0.130960\n",
       "q2    0.110471\n",
       "q3    0.095200\n",
       "q4    0.092446\n",
       "dtype: float64"
      ]
     },
     "execution_count": 150,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Also get the CAGR\n",
    "(df_qsum.iloc[3,:]/df_qsum.iloc[0,:])**(1/3) - 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Further Insights\n",
    "\n",
    "After exploring the data through a Random Forest model, both in feature importance and impact I can now see that a lot of the high correlations seen above were likely to have been artefacts of other correlated features. I.e. the immigrant populations could have been correlated with population density and dwelling structure, and thus by focusing on those two outputs their importance was not as high as originally featured.\n",
    "\n",
    "I am very happy with my evaluation of the Random Forest model, delivering an R2 score of 0.745 is great.\n",
    "\n",
    "By looking at the \"Impacts\" of changing values in these key features, it is interesting to notice the similarity to log probability functions for \"Employed in agriculture, forestry and fishing%\" and \"population density\" (and latitude when the bounds are pushed beyond standard deviation). Once you hit a certain level of people employed in these \"non-urban\" industries, you switch from a -2.5% impact on solar installations to a +4.5% - quite interesting.\n",
    "\n",
    "I'm intrigued by the impact of Latitude on solar installations, as this is contrary to basic science indicating that solar has higher potential at lower latitutdes (closer to the equator) as nicely demonstrated [here](http://profhorn.meteor.wisc.edu/wxwise/radiation/sunangle.html). Whether it is a proxy for government intervention (subsidies and rebates at a state level), an indicator of higher electricty prices in the south, reflective of a more temperate climate meaning the effective yield is higher (due to fewer cloudy days), are all open questions for further exploration.\n",
    "\n",
    "The above is all quite technical, so when it gets to cutting things into quartiles I'd better humanise it all a bit, by finding a representative \"Town\" for each of the quartiles to give a real human feel to what each area looks like."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "q4    141\n",
       "q3    140\n",
       "q2    140\n",
       "q1    135\n",
       "Name: 2016 Quartile, dtype: int64"
      ]
     },
     "execution_count": 175,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pvt['2016 Quartile'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "q1 Junee\n",
      "q2 Lennox Head - Skennars Head\n",
      "q3 West Pennant Hills\n",
      "q4 Lilyfield - Rozelle\n"
     ]
    }
   ],
   "source": [
    "# Get some region names from each of the quartiles\n",
    "# Middle points\n",
    "for q in q_list:\n",
    "    code_ref = pvt[pvt['2016 Quartile']==q].sort_values('2016_x').reset_index().iloc[69,0]\n",
    "    print(q, df[df['CODE'] == code_ref]['LABEL'].iloc[0])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "q1 Helensburgh\n",
      "q2 Wagga Wagga - South\n",
      "q3 Fairfield - East\n",
      "q4 Potts Point - Woolloomooloo\n"
     ]
    }
   ],
   "source": [
    "# Get some region names from each of the quartiles\n",
    "# end points\n",
    "for q in q_list:\n",
    "    code_ref = pvt[pvt['2016 Quartile']==q].sort_values('2016_x').reset_index().iloc[-1,0]\n",
    "    print(q, df[df['CODE'] == code_ref]['LABEL'].iloc[0])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Based on the above, what insights would I share with key stakeholders to the solar industry (energy companies, governments, etc.) when considering future strategies?\n",
    "\n",
    "For those in the business of supplying solar panels: it looks like the best market in Australia is outside the major cities, where population density is far lower and you have fewer apartment type structures. These areas are showing the strongest trends in growth in installations, and already represent around 60% of the total installations in Australia (meaning good business for those looking to leverage future maintenance opportunities).\n",
    "\n",
    "From the view of the public sector, it looks like the market to-date is doing a good job of optimising for the population most likely to benefit from self-purchased solar installations:\n",
    "1. Closer generation to the source means less transmission loss, so a net efficiency overall. \n",
    "2. Non-urban areas also benefit from installations due to the impact of \"remoteness\" on ability for power companies to quickly resolve maintenance issues.\n",
    "\n",
    "Of course, the factor \"Latitude\" plays into this (and how closely it ties to state government incentives) may also be an indicator that the more \"Southern\" governments of New South Wales, Victoria and Tasmania are also doing a good job in further incentivising investment in this particular technology.\n",
    "\n",
    "On the downside, housing trends are heading toward increased population density, and lower levels of home ownership. This means that continued progress to renewable goals will require investment outside of subsidising solar panel installation, and coming up with solutions relevant to and scaleable for urban populations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}