Modeling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predicting Kickstarter project success using ML models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The data cleaning was performed in the Jupyter notebook file: ```'1_Data_Cleaning.ipynb'```, which produces the file ```'Kickstarter_clean.csv'```.\n",
    "\n",
    "This notebook deals with taking the clean data and using machine learning models to predict the outcome of a Kickstarter project."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Importing packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix, classification_report\n",
    "import itertools\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "RSEED = 42"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading cleaned data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('data/Kickstarter_clean.csv', sep=',')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "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>backers_count</th>\n",
       "      <th>category</th>\n",
       "      <th>country</th>\n",
       "      <th>goal</th>\n",
       "      <th>staff_pick</th>\n",
       "      <th>usd_pledged</th>\n",
       "      <th>target</th>\n",
       "      <th>blurb_length</th>\n",
       "      <th>name_length</th>\n",
       "      <th>subcategory</th>\n",
       "      <th>duration_days</th>\n",
       "      <th>year</th>\n",
       "      <th>month</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>21</td>\n",
       "      <td>music</td>\n",
       "      <td>US</td>\n",
       "      <td>200.0</td>\n",
       "      <td>0</td>\n",
       "      <td>802.00000</td>\n",
       "      <td>1</td>\n",
       "      <td>26</td>\n",
       "      <td>4</td>\n",
       "      <td>rock</td>\n",
       "      <td>45</td>\n",
       "      <td>2013</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>97</td>\n",
       "      <td>art</td>\n",
       "      <td>US</td>\n",
       "      <td>400.0</td>\n",
       "      <td>0</td>\n",
       "      <td>2259.00000</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>5</td>\n",
       "      <td>mixed media</td>\n",
       "      <td>20</td>\n",
       "      <td>2019</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>88</td>\n",
       "      <td>photography</td>\n",
       "      <td>US</td>\n",
       "      <td>27224.0</td>\n",
       "      <td>1</td>\n",
       "      <td>29638.00000</td>\n",
       "      <td>1</td>\n",
       "      <td>25</td>\n",
       "      <td>9</td>\n",
       "      <td>photobooks</td>\n",
       "      <td>30</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>193</td>\n",
       "      <td>fashion</td>\n",
       "      <td>IT</td>\n",
       "      <td>45461.0</td>\n",
       "      <td>0</td>\n",
       "      <td>49075.15252</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>5</td>\n",
       "      <td>footwear</td>\n",
       "      <td>41</td>\n",
       "      <td>2018</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>20</td>\n",
       "      <td>technology</td>\n",
       "      <td>US</td>\n",
       "      <td>1000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>549.00000</td>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>4</td>\n",
       "      <td>software</td>\n",
       "      <td>30</td>\n",
       "      <td>2015</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   backers_count     category country     goal  staff_pick  usd_pledged  \\\n",
       "0             21        music      US    200.0           0    802.00000   \n",
       "1             97          art      US    400.0           0   2259.00000   \n",
       "2             88  photography      US  27224.0           1  29638.00000   \n",
       "3            193      fashion      IT  45461.0           0  49075.15252   \n",
       "4             20   technology      US   1000.0           0    549.00000   \n",
       "\n",
       "   target  blurb_length  name_length  subcategory  duration_days  year  month  \n",
       "0       1            26            4         rock             45  2013     12  \n",
       "1       1             9            5  mixed media             20  2019      2  \n",
       "2       1            25            9   photobooks             30  2016     11  \n",
       "3       1            13            5     footwear             41  2018     10  \n",
       "4       0            22            4     software             30  2015      3  "
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Examining the features of the cleaned dataset\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocessing the data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert some columns to categorical\n",
    "df['country'] = df['country'].astype('category')\n",
    "df['category'] = df['category'].astype('category')\n",
    "df['staff_pick'] = df['staff_pick'].astype('category')\n",
    "df['year'] = df['year'].astype('category')\n",
    "df['month'] = df['month'].astype('category')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```backers_count``` and ``usd_pledged`` are quantities that are known at the deadline of the project. It stands to reason that the higher the number of backers and the higher the amount pledged, the more likely a project is to succeed. We want to predict at the launch of the project whether it will be successful or not and the number of backers and amount pledged is not known at the start of a project's launch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining target and feature variables\n",
    "y = df['target']\n",
    "X = df.drop(['target','backers_count','usd_pledged','subcategory'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "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>category</th>\n",
       "      <th>country</th>\n",
       "      <th>goal</th>\n",
       "      <th>staff_pick</th>\n",
       "      <th>blurb_length</th>\n",
       "      <th>name_length</th>\n",
       "      <th>duration_days</th>\n",
       "      <th>year</th>\n",
       "      <th>month</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>music</td>\n",
       "      <td>US</td>\n",
       "      <td>200.0</td>\n",
       "      <td>0</td>\n",
       "      <td>26</td>\n",
       "      <td>4</td>\n",
       "      <td>45</td>\n",
       "      <td>2013</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>art</td>\n",
       "      <td>US</td>\n",
       "      <td>400.0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>5</td>\n",
       "      <td>20</td>\n",
       "      <td>2019</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>photography</td>\n",
       "      <td>US</td>\n",
       "      <td>27224.0</td>\n",
       "      <td>1</td>\n",
       "      <td>25</td>\n",
       "      <td>9</td>\n",
       "      <td>30</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>fashion</td>\n",
       "      <td>IT</td>\n",
       "      <td>45461.0</td>\n",
       "      <td>0</td>\n",
       "      <td>13</td>\n",
       "      <td>5</td>\n",
       "      <td>41</td>\n",
       "      <td>2018</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>technology</td>\n",
       "      <td>US</td>\n",
       "      <td>1000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>22</td>\n",
       "      <td>4</td>\n",
       "      <td>30</td>\n",
       "      <td>2015</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      category country     goal staff_pick  blurb_length  name_length  \\\n",
       "0        music      US    200.0          0            26            4   \n",
       "1          art      US    400.0          0             9            5   \n",
       "2  photography      US  27224.0          1            25            9   \n",
       "3      fashion      IT  45461.0          0            13            5   \n",
       "4   technology      US   1000.0          0            22            4   \n",
       "\n",
       "   duration_days  year month  \n",
       "0             45  2013    12  \n",
       "1             20  2019     2  \n",
       "2             30  2016    11  \n",
       "3             41  2018    10  \n",
       "4             30  2015     3  "
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make copy of features DataFrame for error analysis later\n",
    "Xc = X.copy()\n",
    "Xc_train, Xc_test, yc_train, yc_test = train_test_split(Xc, y, test_size=0.3, stratify=y, random_state=RSEED)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating dummy variables\n",
    "col_dummies = ['category', 'country', 'staff_pick', 'year', 'month']\n",
    "X_dum = pd.get_dummies(X[col_dummies], drop_first=True)\n",
    "X.drop(col_dummies, axis=1, inplace=True)\n",
    "X = pd.concat([X, X_dum], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "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>goal</th>\n",
       "      <th>blurb_length</th>\n",
       "      <th>name_length</th>\n",
       "      <th>duration_days</th>\n",
       "      <th>category_comics</th>\n",
       "      <th>category_crafts</th>\n",
       "      <th>category_dance</th>\n",
       "      <th>category_design</th>\n",
       "      <th>category_fashion</th>\n",
       "      <th>category_film &amp; video</th>\n",
       "      <th>...</th>\n",
       "      <th>month_3</th>\n",
       "      <th>month_4</th>\n",
       "      <th>month_5</th>\n",
       "      <th>month_6</th>\n",
       "      <th>month_7</th>\n",
       "      <th>month_8</th>\n",
       "      <th>month_9</th>\n",
       "      <th>month_10</th>\n",
       "      <th>month_11</th>\n",
       "      <th>month_12</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>200.0</td>\n",
       "      <td>26</td>\n",
       "      <td>4</td>\n",
       "      <td>45</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>400.0</td>\n",
       "      <td>9</td>\n",
       "      <td>5</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>27224.0</td>\n",
       "      <td>25</td>\n",
       "      <td>9</td>\n",
       "      <td>30</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>45461.0</td>\n",
       "      <td>13</td>\n",
       "      <td>5</td>\n",
       "      <td>41</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>22</td>\n",
       "      <td>4</td>\n",
       "      <td>30</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 61 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      goal  blurb_length  name_length  duration_days  category_comics  \\\n",
       "0    200.0            26            4             45                0   \n",
       "1    400.0             9            5             20                0   \n",
       "2  27224.0            25            9             30                0   \n",
       "3  45461.0            13            5             41                0   \n",
       "4   1000.0            22            4             30                0   \n",
       "\n",
       "   category_crafts  category_dance  category_design  category_fashion  \\\n",
       "0                0               0                0                 0   \n",
       "1                0               0                0                 0   \n",
       "2                0               0                0                 0   \n",
       "3                0               0                0                 1   \n",
       "4                0               0                0                 0   \n",
       "\n",
       "   category_film & video  ...  month_3  month_4  month_5  month_6  month_7  \\\n",
       "0                      0  ...        0        0        0        0        0   \n",
       "1                      0  ...        0        0        0        0        0   \n",
       "2                      0  ...        0        0        0        0        0   \n",
       "3                      0  ...        0        0        0        0        0   \n",
       "4                      0  ...        1        0        0        0        0   \n",
       "\n",
       "   month_8  month_9  month_10  month_11  month_12  \n",
       "0        0        0         0         0         1  \n",
       "1        0        0         0         0         0  \n",
       "2        0        0         0         1         0  \n",
       "3        0        0         1         0         0  \n",
       "4        0        0         0         0         0  \n",
       "\n",
       "[5 rows x 61 columns]"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After creating the dummy variables, we can see that the dataset consists of mostly categorical features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split test and train data\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, stratify=y, random_state=RSEED)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Columns that will be scaled: ['goal', 'blurb_length', 'name_length', 'duration_days']\n"
     ]
    }
   ],
   "source": [
    "# Print columns that will be scaled\n",
    "col_scale = list(X.dtypes[(X.dtypes == 'float64') | (X.dtypes == 'int64')].index)\n",
    "print(\"Columns that will be scaled:\", col_scale)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Scaling the data\n",
    "scaler = StandardScaler()\n",
    "X_train_scaled = X_train.copy()\n",
    "X_test_scaled = X_test.copy()\n",
    "X_train_scaled[col_scale] = scaler.fit_transform(X_train[col_scale])\n",
    "X_test_scaled[col_scale] = scaler.transform(X_test[col_scale])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Machine learning models\n",
    "\n",
    "Since we have many categorical features, tree-based models are more suitable for training our model. We also consider a simple logistic regression model. It is not necessary to scale the features for the tree-based models, however scaled data is used for the logistic regression model.\n",
    "\n",
    "The models under consideration are:\n",
    "1. Logistic regression\n",
    "2. Decision tree\n",
    "3. Random forest\n",
    "4. Stochastic Gradient Boost (XGBoost)\n",
    "5. Adaboost"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Logistic regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision on train data:  0.734\n",
      "Accuracy on train data :   0.721\n",
      "Precision on test data:  0.732\n",
      "Accuracy on test data :   0.721\n"
     ]
    }
   ],
   "source": [
    "logreg = LogisticRegression(random_state=RSEED)\n",
    "logreg.fit(X_train_scaled, y_train)\n",
    "\n",
    "y_pred_train = logreg.predict(X_train_scaled)\n",
    "y_pred_test = logreg.predict(X_test_scaled)\n",
    "\n",
    "print(\"Precision on train data: \", round(precision_score(y_train, y_pred_train),3))\n",
    "print(\"Accuracy on train data :  \", round(accuracy_score(y_train, y_pred_train),3))\n",
    "\n",
    "print(\"Precision on test data: \", round(precision_score(y_test, y_pred_test),3))\n",
    "print(\"Accuracy on test data :  \", round(accuracy_score(y_test, y_pred_test),3))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision on train data:  1.0\n",
      "Accuracy on train data :   0.999\n",
      "Precision on test data:  0.698\n",
      "Accuracy on test data :   0.66\n"
     ]
    }
   ],
   "source": [
    "dtree = DecisionTreeClassifier()\n",
    "dtree.fit(X_train, y_train) \n",
    "\n",
    "y_pred_train = dtree.predict(X_train)\n",
    "y_pred_test = dtree.predict(X_test)\n",
    "\n",
    "print(\"Precision on train data: \", round(precision_score(y_train, y_pred_train),3))\n",
    "print(\"Accuracy on train data :  \", round(accuracy_score(y_train, y_pred_train),3))\n",
    "\n",
    "print(\"Precision on test data: \", round(precision_score(y_test, y_pred_test),3))\n",
    "print(\"Accuracy on test data :  \", round(accuracy_score(y_test, y_pred_test),3))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Random forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:    1.4s\n",
      "[Parallel(n_jobs=-1)]: Done 100 out of 100 | elapsed:    3.2s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_features='sqrt', n_jobs=-1, random_state=42,\n",
       "                       verbose=1)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "# Create the model with 100 trees\n",
    "rf = RandomForestClassifier(n_estimators=100, \n",
    "                               random_state=RSEED, \n",
    "                               max_features = 'sqrt',\n",
    "                               n_jobs=-1, verbose = 1)\n",
    "\n",
    "# Fit on training data\n",
    "rf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=8)]: Using backend ThreadingBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=8)]: Done  34 tasks      | elapsed:    0.4s\n",
      "[Parallel(n_jobs=8)]: Done 100 out of 100 | elapsed:    0.9s finished\n",
      "[Parallel(n_jobs=8)]: Using backend ThreadingBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=8)]: Done  34 tasks      | elapsed:    0.2s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision on train data:  0.999\n",
      "Accuracy on train data :   0.999\n",
      "Precision on test data:  0.752\n",
      "Accuracy on test data :   0.736\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=8)]: Done 100 out of 100 | elapsed:    0.5s finished\n"
     ]
    }
   ],
   "source": [
    "y_pred_train = rf.predict(X_train)\n",
    "y_pred_test = rf.predict(X_test)\n",
    "\n",
    "print(\"Precision on train data: \", round(precision_score(y_train, y_pred_train),3))\n",
    "print(\"Accuracy on train data :  \", round(accuracy_score(y_train, y_pred_train),3))\n",
    "\n",
    "print(\"Precision on test data: \", round(precision_score(y_test, y_pred_test),3))\n",
    "print(\"Accuracy on test data :  \", round(accuracy_score(y_test, y_pred_test),3))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Stochastic Gradient Boost (XGBoost)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', callbacks=None,\n",
       "              colsample_bylevel=1, colsample_bynode=1, colsample_bytree=1,\n",
       "              early_stopping_rounds=None, enable_categorical=False,\n",
       "              eval_metric=None, gamma=0, gpu_id=-1, grow_policy='depthwise',\n",
       "              importance_type=None, interaction_constraints='',\n",
       "              learning_rate=0.300000012, max_bin=256, max_cat_to_onehot=4,\n",
       "              max_delta_step=0, max_depth=6, max_leaves=0, min_child_weight=1,\n",
       "              missing=nan, monotone_constraints='()', n_estimators=100,\n",
       "              n_jobs=0, num_parallel_tree=1, predictor='auto', random_state=42,\n",
       "              reg_alpha=0, reg_lambda=1, ...)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb = XGBClassifier(random_state=RSEED)\n",
    "xgb.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision on train data:  0.782\n",
      "Accuracy on train data :   0.775\n",
      "Precision on test data:  0.76\n",
      "Accuracy on test data :   0.7511\n"
     ]
    }
   ],
   "source": [
    "y_pred_train = xgb.predict(X_train)\n",
    "y_pred_test = xgb.predict(X_test)\n",
    "\n",
    "print(\"Precision on train data: \", round(precision_score(y_train, y_pred_train),3))\n",
    "print(\"Accuracy on train data :  \", round(accuracy_score(y_train, y_pred_train),3))\n",
    "\n",
    "print(\"Precision on test data: \", round(precision_score(y_test, y_pred_test),4))\n",
    "print(\"Accuracy on test data :  \", round(accuracy_score(y_test, y_pred_test),4))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6. Adaboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AdaBoostClassifier(base_estimator=DecisionTreeClassifier(), random_state=42)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = AdaBoostClassifier(base_estimator=DecisionTreeClassifier(), random_state=RSEED) \n",
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision on train data:  1.0\n",
      "Accuracy on train data :   0.999\n",
      "Precision on test data:  0.717\n",
      "Accuracy on test data :   0.696\n"
     ]
    }
   ],
   "source": [
    "y_pred_train = clf.predict(X_train)\n",
    "y_pred_test = clf.predict(X_test)\n",
    "\n",
    "print(\"Precision on train data: \", round(precision_score(y_train, y_pred_train),3))\n",
    "print(\"Accuracy on train data :  \", round(accuracy_score(y_train, y_pred_train),3))\n",
    "\n",
    "print(\"Precision on test data: \", round(precision_score(y_test, y_pred_test),3))\n",
    "print(\"Accuracy on test data :  \", round(accuracy_score(y_test, y_pred_test),3))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The decision tree, random forest and Adaboost models showed a significant difference in the precision and accuracy between the train and test data set, which is indicative of overfitting."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Best model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "XGBoost gave the best results \"out of the box\", so we will run a grid search and cross validation using it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 1 candidates, totalling 5 fits\n",
      "[CV] END ...learning_rate=0.4, max_depth=5, n_estimators=100; total time=   5.1s\n",
      "[CV] END ...learning_rate=0.4, max_depth=5, n_estimators=100; total time=   5.0s\n",
      "[CV] END ...learning_rate=0.4, max_depth=5, n_estimators=100; total time=   4.8s\n",
      "[CV] END ...learning_rate=0.4, max_depth=5, n_estimators=100; total time=   5.1s\n",
      "[CV] END ...learning_rate=0.4, max_depth=5, n_estimators=100; total time=   5.3s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5,\n",
       "             estimator=XGBClassifier(base_score=None, booster=None,\n",
       "                                     callbacks=None, colsample_bylevel=None,\n",
       "                                     colsample_bynode=None,\n",
       "                                     colsample_bytree=None,\n",
       "                                     early_stopping_rounds=None,\n",
       "                                     enable_categorical=False, eval_metric=None,\n",
       "                                     gamma=None, gpu_id=None, grow_policy=None,\n",
       "                                     importance_type=None,\n",
       "                                     interaction_constraints=None,\n",
       "                                     learning_rate=None, max_bin=None,\n",
       "                                     max_cat_to_onehot=None,\n",
       "                                     max_delta_step=None, max_depth=None,\n",
       "                                     max_leaves=None, min_child_weight=None,\n",
       "                                     missing=nan, monotone_constraints=None,\n",
       "                                     n_estimators=100, n_jobs=None,\n",
       "                                     num_parallel_tree=None, predictor=None,\n",
       "                                     random_state=42, reg_alpha=None,\n",
       "                                     reg_lambda=None, ...),\n",
       "             param_grid={'learning_rate': [0.4], 'max_depth': [5],\n",
       "                         'n_estimators': [100]},\n",
       "             scoring='precision', verbose=2)"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "estimator = XGBClassifier(random_state=RSEED)\n",
    "\n",
    "param_grid = {\n",
    "    'max_depth': [4, 5, 6, 7, 8],\n",
    "    'n_estimators': [50, 100, 150, 200, 250],\n",
    "    'learning_rate': [0.3, 0.4, 0.5, 0.6]\n",
    "}\n",
    "\n",
    "grid_search = GridSearchCV(\n",
    "    estimator=estimator,\n",
    "    param_grid=param_grid,\n",
    "    scoring = 'precision',\n",
    "    cv = 5,\n",
    "    verbose=2\n",
    ")\n",
    "\n",
    "grid_search.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision on train data:  0.778\n",
      "Accuracy on train data :   0.77\n",
      "Precision on test data:  0.761\n",
      "Accuracy on test data :   0.751\n",
      "Best params:  {'learning_rate': 0.4, 'max_depth': 5, 'n_estimators': 100}\n",
      "\n",
      "\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.67      0.70     22260\n",
      "           1       0.76      0.81      0.79     28434\n",
      "\n",
      "    accuracy                           0.75     50694\n",
      "   macro avg       0.75      0.74      0.74     50694\n",
      "weighted avg       0.75      0.75      0.75     50694\n",
      "\n"
     ]
    }
   ],
   "source": [
    "best_model = grid_search.best_estimator_\n",
    "best_params = grid_search.best_params_\n",
    "\n",
    "y_pred_train = best_model.predict(X_train)\n",
    "y_pred_test = best_model.predict(X_test)\n",
    "\n",
    "print(\"Precision on train data: \", round(precision_score(y_train, y_pred_train),3))\n",
    "print(\"Accuracy on train data :  \", round(accuracy_score(y_train, y_pred_train),3))\n",
    "\n",
    "print(\"Precision on test data: \", round(precision_score(y_test, y_pred_test),3))\n",
    "print(\"Accuracy on test data :  \", round(accuracy_score(y_test, y_pred_test),3))\n",
    "\n",
    "print(\"Best params: \", best_params)\n",
    "print(\"\\n\")\n",
    "print(classification_report(y_test, y_pred_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Confusion matrix on train and test data\n",
    "cm_train = confusion_matrix(y_train, y_pred_train)\n",
    "cm_test = confusion_matrix(y_test, y_pred_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to plot confusion matrix\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        print(\"Normalized confusion matrix\")\n",
    "        print(\"--------\"*5)\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization:')\n",
    "        print(\"--------\"*5)\n",
    "\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=0)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, format(cm[i, j], fmt),\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization:\n",
      "----------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(cm_train, classes=['Fail', 'Success'], normalize=False, title='Confusion matrix for train data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization:\n",
      "----------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(cm_test, classes=['Fail', 'Success'], normalize=False, title='Confusion matrix for test data')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Save false positive (FP) and false negative (FN) rows for later error analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Exploratory Data Analysis will be done on the rows where the predicted values do not match the target values to understand how the model can be improved. A DataFrame with only the observations where the target and predicted value do not match is saved. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "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>category</th>\n",
       "      <th>country</th>\n",
       "      <th>goal</th>\n",
       "      <th>staff_pick</th>\n",
       "      <th>blurb_length</th>\n",
       "      <th>name_length</th>\n",
       "      <th>duration_days</th>\n",
       "      <th>year</th>\n",
       "      <th>month</th>\n",
       "      <th>target</th>\n",
       "      <th>y_pred_train</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>77656</th>\n",
       "      <td>crafts</td>\n",
       "      <td>US</td>\n",
       "      <td>25000.0</td>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>9</td>\n",
       "      <td>30</td>\n",
       "      <td>2015</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108153</th>\n",
       "      <td>dance</td>\n",
       "      <td>US</td>\n",
       "      <td>5000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2017</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41239</th>\n",
       "      <td>technology</td>\n",
       "      <td>US</td>\n",
       "      <td>250.0</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>6</td>\n",
       "      <td>30</td>\n",
       "      <td>2017</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132773</th>\n",
       "      <td>art</td>\n",
       "      <td>US</td>\n",
       "      <td>2533.0</td>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>3</td>\n",
       "      <td>59</td>\n",
       "      <td>2014</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76986</th>\n",
       "      <td>art</td>\n",
       "      <td>US</td>\n",
       "      <td>1000.0</td>\n",
       "      <td>0</td>\n",
       "      <td>21</td>\n",
       "      <td>9</td>\n",
       "      <td>30</td>\n",
       "      <td>2018</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          category country     goal staff_pick  blurb_length  name_length  \\\n",
       "77656       crafts      US  25000.0          1            19            9   \n",
       "108153       dance      US   5000.0          0            24            8   \n",
       "41239   technology      US    250.0          0            25            6   \n",
       "132773         art      US   2533.0          0            19            3   \n",
       "76986          art      US   1000.0          0            21            9   \n",
       "\n",
       "        duration_days  year month  target  y_pred_train  \n",
       "77656              30  2015     6       0             1  \n",
       "108153             30  2017     4       1             0  \n",
       "41239              30  2017     3       0             1  \n",
       "132773             59  2014     1       1             0  \n",
       "76986              30  2018     9       0             1  "
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_train = pd.DataFrame(y_pred_train, columns=['y_pred_train']).set_index(y_train.index)\n",
    "df_error = pd.concat([Xc_train, y_train, y_pred_train], axis=1)\n",
    "df_error = df_error.query('target != y_pred_train')\n",
    "df_error.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(27215, 11)"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_error.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_error.to_csv('data/Kickstarter_error_analysis.csv', sep=',', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Error Analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Failure/success rate by category"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1382.83x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_plot =df.groupby('year')['target'].value_counts(normalize=True).rename('percent').mul(100).reset_index()\n",
    "plt.rcParams['legend.title_fontsize'] = 'large'\n",
    "plt.rcParams['legend.fontsize'] = 'large'\n",
    "g = sns.catplot(x=\"year\", y=\"percent\", data=df_plot, hue=\"target\", kind='bar', height=5, aspect=3.7, palette=sns.color_palette(['#d5815f', '#69b0b1'])).set(title='Success rate by year (actual)')\n",
    "plt.ylim((0, 100))\n",
    "plt.yticks(np.arange(0,101,10),fontsize=20)\n",
    "g.set_xlabels(\"Year\", fontsize=28)\n",
    "g.set_xticklabels(fontsize=20)\n",
    "g.set_ylabels(\"Failure/success rate (%)\", fontsize=28)\n",
    "plt.title('Actual success rate by year', fontsize=28)\n",
    "sns.despine(fig=None, ax=None, top=False, right=False, left=False, bottom=False, offset=None, trim=False)\n",
    "\n",
    "legend = g._legend\n",
    "legend.set_title(\"Outcome\")\n",
    "for t, l in zip(legend.texts,(\"Failure\", \"Success\")):\n",
    "    t.set_text(l)\n",
    "\n",
    "g.fig.subplots_adjust(top=1,right=0.89)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Predicted success rate by year where the prediction was wrong"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1416.35x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_plot = df_error.groupby('year')['y_pred_train'].value_counts(normalize=True).mul(100).rename('percent').reset_index()\n",
    "plt.rcParams['legend.title_fontsize'] = 'large'\n",
    "plt.rcParams['legend.fontsize'] = 'large'\n",
    "g = sns.catplot(x=\"year\", y=\"percent\", data=df_plot, hue=\"y_pred_train\", kind='bar', height=5, aspect=3.7, palette=sns.color_palette(['#d5815f', '#69b0b1'])).set(title='Predicted success rate by year')\n",
    "plt.ylim((0, 100))\n",
    "plt.yticks(np.arange(0,101,10),fontsize=20)\n",
    "g.set_xlabels(\"Year\", fontsize=28)\n",
    "g.set_xticklabels(fontsize=20)\n",
    "g.set_ylabels(\"Failure/success rate (%)\", fontsize=28)\n",
    "plt.title('Predicted success rate by year', fontsize=28)\n",
    "sns.despine(fig=None, ax=None, top=False, right=False, left=False, bottom=False, offset=None, trim=False)\n",
    "\n",
    "legend = g._legend\n",
    "legend.set_title(\"Outcome\")\n",
    "for t, l in zip(legend.texts,(\"Failure\", \"Success\")):\n",
    "    t.set_text(l)\n",
    "\n",
    "g.fig.subplots_adjust(top=1,right=0.89)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.8 ('.venv': venv)",
   "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": "3.9.8"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "e26904dc154732a495317b113324a53c54b86c05343918c18aeac298387fd9fa"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}