LSTM.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import re\n",
    "import ast\n",
    "import glob\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "\n",
    "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, confusion_matrix\n",
    "from sklearn.model_selection import train_test_split, cross_val_score\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.pipeline import Pipeline, make_pipeline\n",
    "\n",
    "\n",
    "from tensorflow.python.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.python.keras.preprocessing.sequence import pad_sequences\n",
    "from tensorflow.python.keras.models import Sequential\n",
    "from tensorflow.python.keras import layers\n",
    "\n",
    "plt.style.use('ggplot')\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_columns', 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "master = pd.read_csv('data/master_cleaned.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/IPython/core/interactiveshell.py:3063: 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": [
    "comments = pd.read_csv('data/comments.csv', usecols=[0,1,2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "dates = pd.read_csv('data/comments_dates.csv', names=['name','url','dates'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 19789 entries, 0 to 19788\n",
      "Data columns (total 3 columns):\n",
      " #   Column      Non-Null Count  Dtype \n",
      "---  ------      --------------  ----- \n",
      " 0   Unnamed: 0  19789 non-null  object\n",
      " 1   url         19789 non-null  object\n",
      " 2   comments    19789 non-null  object\n",
      "dtypes: object(3)\n",
      "memory usage: 463.9+ KB\n"
     ]
    }
   ],
   "source": [
    "comments.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "comments.drop('Unnamed: 0', axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>url</th>\n",
       "      <th>comments</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>https://www.kickstarter.com/projects/117862918...</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>https://www.kickstarter.com/projects/lizardsho...</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>https://www.kickstarter.com/projects/184368767...</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>https://www.kickstarter.com/projects/376746530...</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>https://www.kickstarter.com/projects/116371417...</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 url comments\n",
       "0  https://www.kickstarter.com/projects/117862918...       []\n",
       "1  https://www.kickstarter.com/projects/lizardsho...       []\n",
       "2  https://www.kickstarter.com/projects/184368767...       []\n",
       "3  https://www.kickstarter.com/projects/376746530...       []\n",
       "4  https://www.kickstarter.com/projects/116371417...       []"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "comments.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "comments['comments'][11] = []\n",
    "comments['comments'][12] = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "urls = pd.read_csv('data/comments_urls_copy.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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>Unnamed: 0</th>\n",
       "      <th>id</th>\n",
       "      <th>name</th>\n",
       "      <th>url</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>751419376</td>\n",
       "      <td>Iron Age Kingdoms - First Free To Win Mobile S...</td>\n",
       "      <td>https://www.kickstarter.com/projects/117862918...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>289952460</td>\n",
       "      <td>\"Lizard\" a short film</td>\n",
       "      <td>https://www.kickstarter.com/projects/lizardsho...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>37585826</td>\n",
       "      <td>Total Franchise Football 2016</td>\n",
       "      <td>https://www.kickstarter.com/projects/184368767...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>1727301949</td>\n",
       "      <td>The Trans-American Psychogeographic Literary C...</td>\n",
       "      <td>https://www.kickstarter.com/projects/376746530...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>144196412</td>\n",
       "      <td>The Golden Ticket to the Wonka Factory</td>\n",
       "      <td>https://www.kickstarter.com/projects/116371417...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0          id                                               name  \\\n",
       "0           0   751419376  Iron Age Kingdoms - First Free To Win Mobile S...   \n",
       "1           1   289952460                              \"Lizard\" a short film   \n",
       "2           2    37585826                      Total Franchise Football 2016   \n",
       "3           3  1727301949  The Trans-American Psychogeographic Literary C...   \n",
       "4           4   144196412             The Golden Ticket to the Wonka Factory   \n",
       "\n",
       "                                                 url  \n",
       "0  https://www.kickstarter.com/projects/117862918...  \n",
       "1  https://www.kickstarter.com/projects/lizardsho...  \n",
       "2  https://www.kickstarter.com/projects/184368767...  \n",
       "3  https://www.kickstarter.com/projects/376746530...  \n",
       "4  https://www.kickstarter.com/projects/116371417...  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "urls.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# merge comments table with url tables to get campaign id\n",
    "full = pd.merge(urls, comments, how='right', on= 'url', )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_with_dates = pd.merge(full, dates, on='url')\n",
    "full_with_dates.drop(columns=['Unnamed: 0', 'name_y'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "full_with_dates.rename(columns={'name_x': 'name'}, inplace=True)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Merge with Master"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# full df with comments and dates\n",
    "df = pd.merge(master, full_with_dates, on='name')\n",
    "df.drop(columns=['Unnamed: 0', 'country_displayable_name'], inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(169657, 25)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# master list with outer join, with comments or no comments \n",
    "df_full = pd.merge(master, full_with_dates, on='name', how='left')\n",
    "df_full.drop(columns=['Unnamed: 0','country_displayable_name'], inplace=True)\n",
    "df_full.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols_to_convert = ['deadline', 'launched_at']\n",
    "for c in cols_to_convert:\n",
    "    df[c] = pd.to_datetime(df[c])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols_to_convert = ['deadline', 'launched_at']\n",
    "for c in cols_to_convert:\n",
    "    df_full[c] = pd.to_datetime(df_full[c])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['mid_campaign'] = df['launched_at'] + pd.DateOffset(15)\n",
    "df['mid_campaign'] = df['mid_campaign'].dt.round('d')\n",
    "df['dates'] = df['dates'].map(lambda x: x.split(\"',\"))\n",
    "df['comment_length'] = df['dates'].map(lambda x: len(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract(cat):\n",
    "    x = ast.literal_eval(re.search('({.+})', cat).group(0))\n",
    "    if 'parent_name' not in x.keys():\n",
    "        return x['slug']\n",
    "    else:\n",
    "        return x['parent_name'].lower()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['category_type'] = df['category'].map(lambda x: extract(x))\n",
    "df = df[df['state'].isin(['successful', 'failed'])]\n",
    "df_full['category_type'] = df_full['category'].map(lambda x: extract(x))\n",
    "df_full = df_full[df_full['state'].isin(['successful', 'failed'])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(8427, 27)\n",
      "(169657, 25)\n"
     ]
    }
   ],
   "source": [
    "print(df.shape)\n",
    "print(df_full.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "scrolled": true
   },
   "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>blurb</th>\n",
       "      <th>category</th>\n",
       "      <th>deadline</th>\n",
       "      <th>goal</th>\n",
       "      <th>launched_at</th>\n",
       "      <th>name</th>\n",
       "      <th>pledged</th>\n",
       "      <th>slug</th>\n",
       "      <th>staff_pick</th>\n",
       "      <th>state</th>\n",
       "      <th>urls</th>\n",
       "      <th>launched_year</th>\n",
       "      <th>launched_month</th>\n",
       "      <th>campaign_length</th>\n",
       "      <th>category_type</th>\n",
       "      <th>blurb_length</th>\n",
       "      <th>country_name</th>\n",
       "      <th>scrape_url</th>\n",
       "      <th>complete_url</th>\n",
       "      <th>scrape_urls</th>\n",
       "      <th>id</th>\n",
       "      <th>url</th>\n",
       "      <th>comments</th>\n",
       "      <th>dates</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5</td>\n",
       "      <td>A Free to Win Online Mobile Strategy Game to e...</td>\n",
       "      <td>{\"id\":272,\"name\":\"Mobile Games\",\"slug\":\"games/...</td>\n",
       "      <td>2017-10-09 14:41:03</td>\n",
       "      <td>1000.0</td>\n",
       "      <td>2017-09-25 14:41:03</td>\n",
       "      <td>Iron Age Kingdoms - First Free To Win Mobile S...</td>\n",
       "      <td>105.0</td>\n",
       "      <td>iron-age-kingdoms-first-free-to-win-mobile-str...</td>\n",
       "      <td>False</td>\n",
       "      <td>failed</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2017</td>\n",
       "      <td>9</td>\n",
       "      <td>14</td>\n",
       "      <td>games</td>\n",
       "      <td>24</td>\n",
       "      <td>Canada</td>\n",
       "      <td>https://www.kickstarter.com/projects/117862918...</td>\n",
       "      <td>https://www.kickstarter.com/projects/117862918...</td>\n",
       "      <td>https://www.kickstarter.com/projects/117862918...</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>19</td>\n",
       "      <td>Lizard is a dark comedy and modern western, th...</td>\n",
       "      <td>{\"id\":11,\"name\":\"Film &amp; Video\",\"slug\":\"film &amp; ...</td>\n",
       "      <td>2019-09-21 23:39:48</td>\n",
       "      <td>1100.0</td>\n",
       "      <td>2019-07-23 23:39:48</td>\n",
       "      <td>\"Lizard\" a short film</td>\n",
       "      <td>2358.0</td>\n",
       "      <td>lizard-a-short-film</td>\n",
       "      <td>False</td>\n",
       "      <td>successful</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2019</td>\n",
       "      <td>7</td>\n",
       "      <td>60</td>\n",
       "      <td>film &amp; video</td>\n",
       "      <td>15</td>\n",
       "      <td>the United States</td>\n",
       "      <td>https://www.kickstarter.com/projects/lizardsho...</td>\n",
       "      <td>https://www.kickstarter.com/projects/lizardsho...</td>\n",
       "      <td>https://www.kickstarter.com/projects/lizardsho...</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>5</td>\n",
       "      <td>My friend and I are making an app that has nev...</td>\n",
       "      <td>{\"id\":272,\"name\":\"Mobile Games\",\"slug\":\"games/...</td>\n",
       "      <td>2016-03-22 21:11:17</td>\n",
       "      <td>25000.0</td>\n",
       "      <td>2016-01-22 22:11:17</td>\n",
       "      <td>Total Franchise Football 2016</td>\n",
       "      <td>172.0</td>\n",
       "      <td>total-franchise-football-2016</td>\n",
       "      <td>False</td>\n",
       "      <td>failed</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2016</td>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>games</td>\n",
       "      <td>25</td>\n",
       "      <td>the United States</td>\n",
       "      <td>https://www.kickstarter.com/projects/184368767...</td>\n",
       "      <td>https://www.kickstarter.com/projects/184368767...</td>\n",
       "      <td>https://www.kickstarter.com/projects/184368767...</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>23</td>\n",
       "      <td>More than a bike trip, a state of mind. And a ...</td>\n",
       "      <td>{\"id\":48,\"name\":\"Nonfiction\",\"slug\":\"publishin...</td>\n",
       "      <td>2012-09-25 02:33:06</td>\n",
       "      <td>500.0</td>\n",
       "      <td>2012-09-15 02:33:06</td>\n",
       "      <td>The Trans-American Psychogeographic Literary C...</td>\n",
       "      <td>501.0</td>\n",
       "      <td>the-trans-american-psychogeographic-literary-c...</td>\n",
       "      <td>False</td>\n",
       "      <td>successful</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2012</td>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>publishing</td>\n",
       "      <td>12</td>\n",
       "      <td>the United States</td>\n",
       "      <td>https://www.kickstarter.com/projects/376746530...</td>\n",
       "      <td>https://www.kickstarter.com/projects/376746530...</td>\n",
       "      <td>https://www.kickstarter.com/projects/376746530...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>This is the birth of a revolutionary artist......</td>\n",
       "      <td>{\"id\":39,\"name\":\"Hip-Hop\",\"slug\":\"music/hip-ho...</td>\n",
       "      <td>2017-12-11 02:43:11</td>\n",
       "      <td>2500.0</td>\n",
       "      <td>2017-11-10 02:43:11</td>\n",
       "      <td>The Golden Ticket to the Wonka Factory</td>\n",
       "      <td>0.0</td>\n",
       "      <td>the-golden-ticket-to-the-wonka-factory</td>\n",
       "      <td>False</td>\n",
       "      <td>failed</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2017</td>\n",
       "      <td>11</td>\n",
       "      <td>31</td>\n",
       "      <td>music</td>\n",
       "      <td>23</td>\n",
       "      <td>the United States</td>\n",
       "      <td>https://www.kickstarter.com/projects/116371417...</td>\n",
       "      <td>https://www.kickstarter.com/projects/116371417...</td>\n",
       "      <td>https://www.kickstarter.com/projects/116371417...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   backers_count                                              blurb  \\\n",
       "0              5  A Free to Win Online Mobile Strategy Game to e...   \n",
       "1             19  Lizard is a dark comedy and modern western, th...   \n",
       "2              5  My friend and I are making an app that has nev...   \n",
       "3             23  More than a bike trip, a state of mind. And a ...   \n",
       "4              0  This is the birth of a revolutionary artist......   \n",
       "\n",
       "                                            category            deadline  \\\n",
       "0  {\"id\":272,\"name\":\"Mobile Games\",\"slug\":\"games/... 2017-10-09 14:41:03   \n",
       "1  {\"id\":11,\"name\":\"Film & Video\",\"slug\":\"film & ... 2019-09-21 23:39:48   \n",
       "2  {\"id\":272,\"name\":\"Mobile Games\",\"slug\":\"games/... 2016-03-22 21:11:17   \n",
       "3  {\"id\":48,\"name\":\"Nonfiction\",\"slug\":\"publishin... 2012-09-25 02:33:06   \n",
       "4  {\"id\":39,\"name\":\"Hip-Hop\",\"slug\":\"music/hip-ho... 2017-12-11 02:43:11   \n",
       "\n",
       "      goal         launched_at  \\\n",
       "0   1000.0 2017-09-25 14:41:03   \n",
       "1   1100.0 2019-07-23 23:39:48   \n",
       "2  25000.0 2016-01-22 22:11:17   \n",
       "3    500.0 2012-09-15 02:33:06   \n",
       "4   2500.0 2017-11-10 02:43:11   \n",
       "\n",
       "                                                name  pledged  \\\n",
       "0  Iron Age Kingdoms - First Free To Win Mobile S...    105.0   \n",
       "1                              \"Lizard\" a short film   2358.0   \n",
       "2                      Total Franchise Football 2016    172.0   \n",
       "3  The Trans-American Psychogeographic Literary C...    501.0   \n",
       "4             The Golden Ticket to the Wonka Factory      0.0   \n",
       "\n",
       "                                                slug  staff_pick       state  \\\n",
       "0  iron-age-kingdoms-first-free-to-win-mobile-str...       False      failed   \n",
       "1                                lizard-a-short-film       False  successful   \n",
       "2                      total-franchise-football-2016       False      failed   \n",
       "3  the-trans-american-psychogeographic-literary-c...       False  successful   \n",
       "4             the-golden-ticket-to-the-wonka-factory       False      failed   \n",
       "\n",
       "                                                urls  launched_year  \\\n",
       "0  {\"web\":{\"project\":\"https://www.kickstarter.com...           2017   \n",
       "1  {\"web\":{\"project\":\"https://www.kickstarter.com...           2019   \n",
       "2  {\"web\":{\"project\":\"https://www.kickstarter.com...           2016   \n",
       "3  {\"web\":{\"project\":\"https://www.kickstarter.com...           2012   \n",
       "4  {\"web\":{\"project\":\"https://www.kickstarter.com...           2017   \n",
       "\n",
       "   launched_month  campaign_length category_type  blurb_length  \\\n",
       "0               9               14         games            24   \n",
       "1               7               60  film & video            15   \n",
       "2               1               60         games            25   \n",
       "3               9               10    publishing            12   \n",
       "4              11               31         music            23   \n",
       "\n",
       "        country_name                                         scrape_url  \\\n",
       "0             Canada  https://www.kickstarter.com/projects/117862918...   \n",
       "1  the United States  https://www.kickstarter.com/projects/lizardsho...   \n",
       "2  the United States  https://www.kickstarter.com/projects/184368767...   \n",
       "3  the United States  https://www.kickstarter.com/projects/376746530...   \n",
       "4  the United States  https://www.kickstarter.com/projects/116371417...   \n",
       "\n",
       "                                        complete_url  \\\n",
       "0  https://www.kickstarter.com/projects/117862918...   \n",
       "1  https://www.kickstarter.com/projects/lizardsho...   \n",
       "2  https://www.kickstarter.com/projects/184368767...   \n",
       "3  https://www.kickstarter.com/projects/376746530...   \n",
       "4  https://www.kickstarter.com/projects/116371417...   \n",
       "\n",
       "                                         scrape_urls  id  url comments dates  \n",
       "0  https://www.kickstarter.com/projects/117862918... NaN  NaN      NaN   NaN  \n",
       "1  https://www.kickstarter.com/projects/lizardsho... NaN  NaN      NaN   NaN  \n",
       "2  https://www.kickstarter.com/projects/184368767... NaN  NaN      NaN   NaN  \n",
       "3  https://www.kickstarter.com/projects/376746530... NaN  NaN      NaN   NaN  \n",
       "4  https://www.kickstarter.com/projects/116371417... NaN  NaN      NaN   NaN  "
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_full.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_date(x):\n",
    "    lst = []\n",
    "    for i in x:\n",
    "        i = i.replace(\"[\", '').replace(\"]\", '')\n",
    "        d = pd.to_datetime(i[:-4])\n",
    "        lst.append(d)\n",
    "    return lst"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['dates'] = df['dates'].map(lambda x: to_date(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def date_subtract(df):\n",
    "    lst = []\n",
    "    for index, row in df.iterrows():\n",
    "        row_lst =[]\n",
    "        r = row['dates']\n",
    "        for i in r:\n",
    "            if i < row['mid_campaign']:\n",
    "                row_lst.append(i)\n",
    "        lst.append(row_lst)\n",
    "    return lst "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "def real_comments(df):\n",
    "    lst = []\n",
    "    for index, row in df.iterrows():\n",
    "        r = eval(row['comments'])\n",
    "        a = row['pre_mid_campaign_dates_length']\n",
    "        if a > 0:\n",
    "            lst.append(r[-a:])\n",
    "        else:\n",
    "            lst.append([])\n",
    "    return lst"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "def have_comments(x):\n",
    "    if x > 0:\n",
    "        return 1\n",
    "    else:\n",
    "        return 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['pre_mid_campaign_dates'] = date_subtract(df)\n",
    "df['pre_mid_campaign_dates_length'] = df['pre_mid_campaign_dates'].apply(lambda x: len(x))\n",
    "df['have_comments'] = df['pre_mid_campaign_dates_length'].apply(lambda x: have_comments(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['pre_mid_campaign_comments'] = real_comments(df)"
   ]
  },
  {
   "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>backers_count</th>\n",
       "      <th>blurb</th>\n",
       "      <th>category</th>\n",
       "      <th>deadline</th>\n",
       "      <th>goal</th>\n",
       "      <th>launched_at</th>\n",
       "      <th>name</th>\n",
       "      <th>pledged</th>\n",
       "      <th>slug</th>\n",
       "      <th>staff_pick</th>\n",
       "      <th>state</th>\n",
       "      <th>urls</th>\n",
       "      <th>launched_year</th>\n",
       "      <th>launched_month</th>\n",
       "      <th>campaign_length</th>\n",
       "      <th>category_type</th>\n",
       "      <th>blurb_length</th>\n",
       "      <th>country_name</th>\n",
       "      <th>scrape_url</th>\n",
       "      <th>complete_url</th>\n",
       "      <th>scrape_urls</th>\n",
       "      <th>id</th>\n",
       "      <th>url</th>\n",
       "      <th>comments</th>\n",
       "      <th>dates</th>\n",
       "      <th>mid_campaign</th>\n",
       "      <th>comment_length</th>\n",
       "      <th>pre_mid_campaign_dates</th>\n",
       "      <th>pre_mid_campaign_dates_length</th>\n",
       "      <th>have_comments</th>\n",
       "      <th>pre_mid_campaign_comments</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>152</td>\n",
       "      <td>A photography and poetry book featuring contem...</td>\n",
       "      <td>{\"id\":278,\"name\":\"People\",\"slug\":\"photography/...</td>\n",
       "      <td>2017-05-28 17:00:00</td>\n",
       "      <td>15000.0</td>\n",
       "      <td>2017-05-03 16:57:18</td>\n",
       "      <td>Burma Storybook</td>\n",
       "      <td>15777.00</td>\n",
       "      <td>burma-storybook</td>\n",
       "      <td>True</td>\n",
       "      <td>successful</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2017</td>\n",
       "      <td>5</td>\n",
       "      <td>25</td>\n",
       "      <td>photography</td>\n",
       "      <td>13</td>\n",
       "      <td>the Netherlands</td>\n",
       "      <td>https://www.kickstarter.com/projects/618155130...</td>\n",
       "      <td>https://www.kickstarter.com/projects/618155130...</td>\n",
       "      <td>https://www.kickstarter.com/projects/618155130...</td>\n",
       "      <td>1.769794e+09</td>\n",
       "      <td>https://www.kickstarter.com/projects/618155130...</td>\n",
       "      <td>['Oh! My! The book has arrived and I am speech...</td>\n",
       "      <td>[2017-06-30 10:52:00, 2017-05-26 03:53:00]</td>\n",
       "      <td>2017-05-19</td>\n",
       "      <td>2</td>\n",
       "      <td>[]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>116</td>\n",
       "      <td>100 sumie ink originals by David Mack on top o...</td>\n",
       "      <td>{\"id\":276,\"name\":\"Fine Art\",\"slug\":\"photograph...</td>\n",
       "      <td>2018-07-29 21:05:01</td>\n",
       "      <td>8000.0</td>\n",
       "      <td>2018-06-29 21:05:01</td>\n",
       "      <td>The Alchemy of Collaboration</td>\n",
       "      <td>16706.00</td>\n",
       "      <td>the-alchemy-of-collaboration</td>\n",
       "      <td>False</td>\n",
       "      <td>successful</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2018</td>\n",
       "      <td>6</td>\n",
       "      <td>30</td>\n",
       "      <td>photography</td>\n",
       "      <td>21</td>\n",
       "      <td>the United States</td>\n",
       "      <td>https://www.kickstarter.com/projects/punchgrap...</td>\n",
       "      <td>https://www.kickstarter.com/projects/punchgrap...</td>\n",
       "      <td>https://www.kickstarter.com/projects/punchgrap...</td>\n",
       "      <td>5.263051e+08</td>\n",
       "      <td>https://www.kickstarter.com/projects/punchgrap...</td>\n",
       "      <td>[\"I know you read the messages. I've been chas...</td>\n",
       "      <td>[2020-06-01 10:55:00, 2020-05-21 09:46:00, 202...</td>\n",
       "      <td>2018-07-15</td>\n",
       "      <td>21</td>\n",
       "      <td>[]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>50</td>\n",
       "      <td>8 artists explore the collision of art, the en...</td>\n",
       "      <td>{\"id\":54,\"name\":\"Mixed Media\",\"slug\":\"art/mixe...</td>\n",
       "      <td>2012-06-04 03:59:00</td>\n",
       "      <td>3300.0</td>\n",
       "      <td>2012-05-03 12:13:56</td>\n",
       "      <td>Frack This? - The Wyoming Artist Expedition</td>\n",
       "      <td>3805.00</td>\n",
       "      <td>frack-this-the-wyoming-artist-expedition</td>\n",
       "      <td>False</td>\n",
       "      <td>successful</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2012</td>\n",
       "      <td>5</td>\n",
       "      <td>32</td>\n",
       "      <td>art</td>\n",
       "      <td>16</td>\n",
       "      <td>the United States</td>\n",
       "      <td>https://www.kickstarter.com/projects/331530956...</td>\n",
       "      <td>https://www.kickstarter.com/projects/331530956...</td>\n",
       "      <td>https://www.kickstarter.com/projects/331530956...</td>\n",
       "      <td>1.103726e+09</td>\n",
       "      <td>https://www.kickstarter.com/projects/331530956...</td>\n",
       "      <td>['To the Artists:', 'I encourage each artist t...</td>\n",
       "      <td>[2012-05-30 15:37:00, 2012-05-14 14:46:00, 201...</td>\n",
       "      <td>2012-05-19</td>\n",
       "      <td>4</td>\n",
       "      <td>[2012-05-14 14:46:00, 2012-05-12 09:48:00, 201...</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>[Lori, the kids will want you and Uncle Michae...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>8</td>\n",
       "      <td>Cookietownworld is an interactive website desi...</td>\n",
       "      <td>{\"id\":272,\"name\":\"Mobile Games\",\"slug\":\"games/...</td>\n",
       "      <td>2016-08-07 16:15:57</td>\n",
       "      <td>95000.0</td>\n",
       "      <td>2016-06-08 16:15:57</td>\n",
       "      <td>Cookietownworld</td>\n",
       "      <td>3229.00</td>\n",
       "      <td>cookietownworld</td>\n",
       "      <td>False</td>\n",
       "      <td>failed</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2016</td>\n",
       "      <td>6</td>\n",
       "      <td>60</td>\n",
       "      <td>games</td>\n",
       "      <td>13</td>\n",
       "      <td>the United States</td>\n",
       "      <td>https://www.kickstarter.com/projects/cookietow...</td>\n",
       "      <td>https://www.kickstarter.com/projects/cookietow...</td>\n",
       "      <td>https://www.kickstarter.com/projects/cookietow...</td>\n",
       "      <td>2.045995e+09</td>\n",
       "      <td>https://www.kickstarter.com/projects/cookietow...</td>\n",
       "      <td>['I think Cookietownworld is an awesome idea. ...</td>\n",
       "      <td>[2016-06-23 08:55:00]</td>\n",
       "      <td>2016-06-24</td>\n",
       "      <td>1</td>\n",
       "      <td>[2016-06-23 08:55:00]</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>[I think Cookietownworld is an awesome idea. I...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>59</td>\n",
       "      <td>Explore a magical world full of monsters and d...</td>\n",
       "      <td>{\"id\":272,\"name\":\"Mobile Games\",\"slug\":\"games/...</td>\n",
       "      <td>2017-12-22 20:25:20</td>\n",
       "      <td>15000.0</td>\n",
       "      <td>2017-11-22 20:25:20</td>\n",
       "      <td>Dunpets Colors: Monster-catching RPG</td>\n",
       "      <td>4310.75</td>\n",
       "      <td>dunpets-colors-dungeon-crawl-and-pets</td>\n",
       "      <td>True</td>\n",
       "      <td>failed</td>\n",
       "      <td>{\"web\":{\"project\":\"https://www.kickstarter.com...</td>\n",
       "      <td>2017</td>\n",
       "      <td>11</td>\n",
       "      <td>30</td>\n",
       "      <td>games</td>\n",
       "      <td>22</td>\n",
       "      <td>Spain</td>\n",
       "      <td>https://www.kickstarter.com/projects/gugamesde...</td>\n",
       "      <td>https://www.kickstarter.com/projects/gugamesde...</td>\n",
       "      <td>https://www.kickstarter.com/projects/gugamesde...</td>\n",
       "      <td>1.228293e+09</td>\n",
       "      <td>https://www.kickstarter.com/projects/gugamesde...</td>\n",
       "      <td>['Any updates about the premier? :)', 'Any pro...</td>\n",
       "      <td>[2018-04-17 18:49:00, 2018-01-22 01:59:00, 201...</td>\n",
       "      <td>2017-12-08</td>\n",
       "      <td>18</td>\n",
       "      <td>[2017-12-03 12:14:00, 2017-12-03 07:10:00, 201...</td>\n",
       "      <td>11</td>\n",
       "      <td>1</td>\n",
       "      <td>[Any progress update on the premiere?, Sorry t...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   backers_count                                              blurb  \\\n",
       "0            152  A photography and poetry book featuring contem...   \n",
       "1            116  100 sumie ink originals by David Mack on top o...   \n",
       "2             50  8 artists explore the collision of art, the en...   \n",
       "3              8  Cookietownworld is an interactive website desi...   \n",
       "4             59  Explore a magical world full of monsters and d...   \n",
       "\n",
       "                                            category            deadline  \\\n",
       "0  {\"id\":278,\"name\":\"People\",\"slug\":\"photography/... 2017-05-28 17:00:00   \n",
       "1  {\"id\":276,\"name\":\"Fine Art\",\"slug\":\"photograph... 2018-07-29 21:05:01   \n",
       "2  {\"id\":54,\"name\":\"Mixed Media\",\"slug\":\"art/mixe... 2012-06-04 03:59:00   \n",
       "3  {\"id\":272,\"name\":\"Mobile Games\",\"slug\":\"games/... 2016-08-07 16:15:57   \n",
       "4  {\"id\":272,\"name\":\"Mobile Games\",\"slug\":\"games/... 2017-12-22 20:25:20   \n",
       "\n",
       "      goal         launched_at                                         name  \\\n",
       "0  15000.0 2017-05-03 16:57:18                              Burma Storybook   \n",
       "1   8000.0 2018-06-29 21:05:01                 The Alchemy of Collaboration   \n",
       "2   3300.0 2012-05-03 12:13:56  Frack This? - The Wyoming Artist Expedition   \n",
       "3  95000.0 2016-06-08 16:15:57                              Cookietownworld   \n",
       "4  15000.0 2017-11-22 20:25:20         Dunpets Colors: Monster-catching RPG   \n",
       "\n",
       "    pledged                                      slug  staff_pick       state  \\\n",
       "0  15777.00                           burma-storybook        True  successful   \n",
       "1  16706.00              the-alchemy-of-collaboration       False  successful   \n",
       "2   3805.00  frack-this-the-wyoming-artist-expedition       False  successful   \n",
       "3   3229.00                           cookietownworld       False      failed   \n",
       "4   4310.75     dunpets-colors-dungeon-crawl-and-pets        True      failed   \n",
       "\n",
       "                                                urls  launched_year  \\\n",
       "0  {\"web\":{\"project\":\"https://www.kickstarter.com...           2017   \n",
       "1  {\"web\":{\"project\":\"https://www.kickstarter.com...           2018   \n",
       "2  {\"web\":{\"project\":\"https://www.kickstarter.com...           2012   \n",
       "3  {\"web\":{\"project\":\"https://www.kickstarter.com...           2016   \n",
       "4  {\"web\":{\"project\":\"https://www.kickstarter.com...           2017   \n",
       "\n",
       "   launched_month  campaign_length category_type  blurb_length  \\\n",
       "0               5               25   photography            13   \n",
       "1               6               30   photography            21   \n",
       "2               5               32           art            16   \n",
       "3               6               60         games            13   \n",
       "4              11               30         games            22   \n",
       "\n",
       "        country_name                                         scrape_url  \\\n",
       "0    the Netherlands  https://www.kickstarter.com/projects/618155130...   \n",
       "1  the United States  https://www.kickstarter.com/projects/punchgrap...   \n",
       "2  the United States  https://www.kickstarter.com/projects/331530956...   \n",
       "3  the United States  https://www.kickstarter.com/projects/cookietow...   \n",
       "4              Spain  https://www.kickstarter.com/projects/gugamesde...   \n",
       "\n",
       "                                        complete_url  \\\n",
       "0  https://www.kickstarter.com/projects/618155130...   \n",
       "1  https://www.kickstarter.com/projects/punchgrap...   \n",
       "2  https://www.kickstarter.com/projects/331530956...   \n",
       "3  https://www.kickstarter.com/projects/cookietow...   \n",
       "4  https://www.kickstarter.com/projects/gugamesde...   \n",
       "\n",
       "                                         scrape_urls            id  \\\n",
       "0  https://www.kickstarter.com/projects/618155130...  1.769794e+09   \n",
       "1  https://www.kickstarter.com/projects/punchgrap...  5.263051e+08   \n",
       "2  https://www.kickstarter.com/projects/331530956...  1.103726e+09   \n",
       "3  https://www.kickstarter.com/projects/cookietow...  2.045995e+09   \n",
       "4  https://www.kickstarter.com/projects/gugamesde...  1.228293e+09   \n",
       "\n",
       "                                                 url  \\\n",
       "0  https://www.kickstarter.com/projects/618155130...   \n",
       "1  https://www.kickstarter.com/projects/punchgrap...   \n",
       "2  https://www.kickstarter.com/projects/331530956...   \n",
       "3  https://www.kickstarter.com/projects/cookietow...   \n",
       "4  https://www.kickstarter.com/projects/gugamesde...   \n",
       "\n",
       "                                            comments  \\\n",
       "0  ['Oh! My! The book has arrived and I am speech...   \n",
       "1  [\"I know you read the messages. I've been chas...   \n",
       "2  ['To the Artists:', 'I encourage each artist t...   \n",
       "3  ['I think Cookietownworld is an awesome idea. ...   \n",
       "4  ['Any updates about the premier? :)', 'Any pro...   \n",
       "\n",
       "                                               dates mid_campaign  \\\n",
       "0         [2017-06-30 10:52:00, 2017-05-26 03:53:00]   2017-05-19   \n",
       "1  [2020-06-01 10:55:00, 2020-05-21 09:46:00, 202...   2018-07-15   \n",
       "2  [2012-05-30 15:37:00, 2012-05-14 14:46:00, 201...   2012-05-19   \n",
       "3                              [2016-06-23 08:55:00]   2016-06-24   \n",
       "4  [2018-04-17 18:49:00, 2018-01-22 01:59:00, 201...   2017-12-08   \n",
       "\n",
       "   comment_length                             pre_mid_campaign_dates  \\\n",
       "0               2                                                 []   \n",
       "1              21                                                 []   \n",
       "2               4  [2012-05-14 14:46:00, 2012-05-12 09:48:00, 201...   \n",
       "3               1                              [2016-06-23 08:55:00]   \n",
       "4              18  [2017-12-03 12:14:00, 2017-12-03 07:10:00, 201...   \n",
       "\n",
       "   pre_mid_campaign_dates_length  have_comments  \\\n",
       "0                              0              0   \n",
       "1                              0              0   \n",
       "2                              3              1   \n",
       "3                              1              1   \n",
       "4                             11              1   \n",
       "\n",
       "                           pre_mid_campaign_comments  \n",
       "0                                                 []  \n",
       "1                                                 []  \n",
       "2  [Lori, the kids will want you and Uncle Michae...  \n",
       "3  [I think Cookietownworld is an awesome idea. I...  \n",
       "4  [Any progress update on the premiere?, Sorry t...  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Lori, the kids will want you and Uncle Michael to sign their card!  Enjoy your trip.',\n",
       " 'My $25 pledge was on behalf of Nina Gibbons, who donated to support this project.  Thanks so much for your donation.',\n",
       " 'Hope you guys have a great trip! My only request-my brother and Lori have to sign my book!']"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eval(df['comments'][2])[-3:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_filtered = df[['name', 'have_comments']].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_master = pd.merge(df_full, df_filtered, on='name', how='left')\n",
    "df_master = df_master[df_master['state'].isin(['successful', 'failed'])]\n",
    "df_master['category_type'] = df_master['category'].map(lambda x: extract(x))\n",
    "df_master['have_comments'].fillna(0, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    165838\n",
       "1.0      4411\n",
       "Name: have_comments, dtype: int64"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_master['have_comments'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## One Hot Encoding "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_master_ohe = pd.get_dummies(df_master[['category_type', 'country_name', 'staff_pick']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.merge(df_master, df_master_ohe, left_index=True, right_index=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "good_data = data[data['state'].isin(['successful', 'failed'])]\n",
    "live = data[data['state'].isin(['live'])]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "good_data['state'] = good_data['state'].replace(to_replace=['successful', 'failed'], value=[1,0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_data = good_data.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = model_data.drop(['backers_count','blurb', 'category', 'deadline', 'launched_at', 'urls', 'scrape_url', 'complete_url',\n",
    "                     'scrape_urls', 'name', 'pledged', 'slug', 'state', 'category_type', 'country_name',\n",
    "                    'staff_pick_y', 'id', 'url', 'comments', 'dates'],\n",
    "                   axis=1)\n",
    "y = model_data['state']\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scraped Campaigns Only\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_filtered = df[['name', 'have_comments']].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_data_ = pd.merge(full, df_filtered, on='name', how='left')\n",
    "model_data_.fillna(0,inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_data1 = pd.merge(data, model_data_, on='name', how='right')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_data1.fillna(0, inplace=True)\n",
    "model_data1['state'] = model_data1['state'].replace(to_replace=['successful', 'failed'], value=[1,0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(26779, 57)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_data1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    18778\n",
       "1.0     8001\n",
       "Name: have_comments_y, dtype: int64"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_data1.have_comments_y.value_counts()"
   ]
  },
  {
   "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></th>\n",
       "      <th>backers_count</th>\n",
       "      <th>blurb</th>\n",
       "      <th>category</th>\n",
       "      <th>deadline</th>\n",
       "      <th>goal</th>\n",
       "      <th>launched_at</th>\n",
       "      <th>name</th>\n",
       "      <th>pledged</th>\n",
       "      <th>slug</th>\n",
       "      <th>staff_pick_x</th>\n",
       "      <th>urls</th>\n",
       "      <th>launched_year</th>\n",
       "      <th>launched_month</th>\n",
       "      <th>campaign_length</th>\n",
       "      <th>category_type</th>\n",
       "      <th>blurb_length</th>\n",
       "      <th>country_name</th>\n",
       "      <th>scrape_url</th>\n",
       "      <th>complete_url</th>\n",
       "      <th>scrape_urls</th>\n",
       "      <th>id_x</th>\n",
       "      <th>url_x</th>\n",
       "      <th>comments_x</th>\n",
       "      <th>dates</th>\n",
       "      <th>have_comments_x</th>\n",
       "      <th>...</th>\n",
       "      <th>category_type_design</th>\n",
       "      <th>category_type_fashion</th>\n",
       "      <th>category_type_film &amp; video</th>\n",
       "      <th>category_type_food</th>\n",
       "      <th>category_type_games</th>\n",
       "      <th>category_type_journalism</th>\n",
       "      <th>category_type_music</th>\n",
       "      <th>category_type_photography</th>\n",
       "      <th>category_type_publishing</th>\n",
       "      <th>category_type_technology</th>\n",
       "      <th>category_type_theater</th>\n",
       "      <th>country_name_Australia</th>\n",
       "      <th>country_name_Canada</th>\n",
       "      <th>country_name_France</th>\n",
       "      <th>country_name_Germany</th>\n",
       "      <th>country_name_Italy</th>\n",
       "      <th>country_name_Mexico</th>\n",
       "      <th>country_name_Spain</th>\n",
       "      <th>country_name_other</th>\n",
       "      <th>country_name_the Netherlands</th>\n",
       "      <th>country_name_the United States</th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>id_y</th>\n",
       "      <th>url_y</th>\n",
       "      <th>comments_y</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>have_comments_y</th>\n",
       "      <th>state</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",
       "      <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",
       "      <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=\"2\" valign=\"top\">0.0</th>\n",
       "      <th>0</th>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>...</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "      <td>8291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>...</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "      <td>10487</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">1.0</th>\n",
       "      <th>0</th>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>...</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "      <td>1550</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>...</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "      <td>6451</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4 rows × 55 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                       backers_count  blurb  category  deadline   goal  \\\n",
       "have_comments_y state                                                    \n",
       "0.0             0               8291   8291      8291      8291   8291   \n",
       "                1              10487  10487     10487     10487  10487   \n",
       "1.0             0               1550   1550      1550      1550   1550   \n",
       "                1               6451   6451      6451      6451   6451   \n",
       "\n",
       "                       launched_at   name  pledged   slug  staff_pick_x  \\\n",
       "have_comments_y state                                                     \n",
       "0.0             0             8291   8291     8291   8291          8291   \n",
       "                1            10487  10487    10487  10487         10487   \n",
       "1.0             0             1550   1550     1550   1550          1550   \n",
       "                1             6451   6451     6451   6451          6451   \n",
       "\n",
       "                        urls  launched_year  launched_month  campaign_length  \\\n",
       "have_comments_y state                                                          \n",
       "0.0             0       8291           8291            8291             8291   \n",
       "                1      10487          10487           10487            10487   \n",
       "1.0             0       1550           1550            1550             1550   \n",
       "                1       6451           6451            6451             6451   \n",
       "\n",
       "                       category_type  blurb_length  country_name  scrape_url  \\\n",
       "have_comments_y state                                                          \n",
       "0.0             0               8291          8291          8291        8291   \n",
       "                1              10487         10487         10487       10487   \n",
       "1.0             0               1550          1550          1550        1550   \n",
       "                1               6451          6451          6451        6451   \n",
       "\n",
       "                       complete_url  scrape_urls   id_x  url_x  comments_x  \\\n",
       "have_comments_y state                                                        \n",
       "0.0             0              8291         8291   8291   8291        8291   \n",
       "                1             10487        10487  10487  10487       10487   \n",
       "1.0             0              1550         1550   1550   1550        1550   \n",
       "                1              6451         6451   6451   6451        6451   \n",
       "\n",
       "                       dates  have_comments_x  ...  category_type_design  \\\n",
       "have_comments_y state                          ...                         \n",
       "0.0             0       8291             8291  ...                  8291   \n",
       "                1      10487            10487  ...                 10487   \n",
       "1.0             0       1550             1550  ...                  1550   \n",
       "                1       6451             6451  ...                  6451   \n",
       "\n",
       "                       category_type_fashion  category_type_film & video  \\\n",
       "have_comments_y state                                                      \n",
       "0.0             0                       8291                        8291   \n",
       "                1                      10487                       10487   \n",
       "1.0             0                       1550                        1550   \n",
       "                1                       6451                        6451   \n",
       "\n",
       "                       category_type_food  category_type_games  \\\n",
       "have_comments_y state                                            \n",
       "0.0             0                    8291                 8291   \n",
       "                1                   10487                10487   \n",
       "1.0             0                    1550                 1550   \n",
       "                1                    6451                 6451   \n",
       "\n",
       "                       category_type_journalism  category_type_music  \\\n",
       "have_comments_y state                                                  \n",
       "0.0             0                          8291                 8291   \n",
       "                1                         10487                10487   \n",
       "1.0             0                          1550                 1550   \n",
       "                1                          6451                 6451   \n",
       "\n",
       "                       category_type_photography  category_type_publishing  \\\n",
       "have_comments_y state                                                        \n",
       "0.0             0                           8291                      8291   \n",
       "                1                          10487                     10487   \n",
       "1.0             0                           1550                      1550   \n",
       "                1                           6451                      6451   \n",
       "\n",
       "                       category_type_technology  category_type_theater  \\\n",
       "have_comments_y state                                                    \n",
       "0.0             0                          8291                   8291   \n",
       "                1                         10487                  10487   \n",
       "1.0             0                          1550                   1550   \n",
       "                1                          6451                   6451   \n",
       "\n",
       "                       country_name_Australia  country_name_Canada  \\\n",
       "have_comments_y state                                                \n",
       "0.0             0                        8291                 8291   \n",
       "                1                       10487                10487   \n",
       "1.0             0                        1550                 1550   \n",
       "                1                        6451                 6451   \n",
       "\n",
       "                       country_name_France  country_name_Germany  \\\n",
       "have_comments_y state                                              \n",
       "0.0             0                     8291                  8291   \n",
       "                1                    10487                 10487   \n",
       "1.0             0                     1550                  1550   \n",
       "                1                     6451                  6451   \n",
       "\n",
       "                       country_name_Italy  country_name_Mexico  \\\n",
       "have_comments_y state                                            \n",
       "0.0             0                    8291                 8291   \n",
       "                1                   10487                10487   \n",
       "1.0             0                    1550                 1550   \n",
       "                1                    6451                 6451   \n",
       "\n",
       "                       country_name_Spain  country_name_other  \\\n",
       "have_comments_y state                                           \n",
       "0.0             0                    8291                8291   \n",
       "                1                   10487               10487   \n",
       "1.0             0                    1550                1550   \n",
       "                1                    6451                6451   \n",
       "\n",
       "                       country_name_the Netherlands  \\\n",
       "have_comments_y state                                 \n",
       "0.0             0                              8291   \n",
       "                1                             10487   \n",
       "1.0             0                              1550   \n",
       "                1                              6451   \n",
       "\n",
       "                       country_name_the United States  Unnamed: 0   id_y  \\\n",
       "have_comments_y state                                                      \n",
       "0.0             0                                8291        8291   8291   \n",
       "                1                               10487       10487  10487   \n",
       "1.0             0                                1550        1550   1550   \n",
       "                1                                6451        6451   6451   \n",
       "\n",
       "                       url_y  comments_y  \n",
       "have_comments_y state                     \n",
       "0.0             0       8291        8291  \n",
       "                1      10487       10487  \n",
       "1.0             0       1550        1550  \n",
       "                1       6451        6451  \n",
       "\n",
       "[4 rows x 55 columns]"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_data1.groupby(['have_comments_y', 'state']).count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "bar_width = 0.25\n",
    "success = [10487, 6451]\n",
    "failed = [8291, 1550]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "r1 = np.arange(len(success))\n",
    "r2 = [x + bar_width for x in r1]\n",
    "\n",
    "plt.bar(r1, success, color='darkorange', width=bar_width, edgecolor='white', label='success')\n",
    "plt.bar(r2, failed, color='green', width=bar_width, edgecolor='white', label='failed')\n",
    "plt.xlabel('Mid-Campaign Comments', fontsize=16)\n",
    "plt.ylabel('Count', fontsize=16)\n",
    "plt.xticks([r + bar_width*0.5 for r in range(len(success))], ['No Comments', 'Have Comments'], fontsize=12)\n",
    "plt.title('Status Group', fontsize=18)\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "X1 = model_data1.drop(['backers_count','blurb', 'category', 'deadline', 'urls', 'scrape_url', 'complete_url', 'scrape_urls',\n",
    "                     'launched_at', 'name', 'pledged', 'slug', 'state', 'category_type', 'country_name', 'id_x', 'url_x',\n",
    "                       'comments_x', 'dates', 'staff_pick_y', 'Unnamed: 0', 'id_y', 'url_y', 'comments_y', 'have_comments_y'],\n",
    "                   axis=1)\n",
    "y1 = model_data1['state']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "X1_old = model_data1.drop(['backers_count','blurb', 'category', 'deadline', 'urls', 'scrape_url', 'complete_url', 'scrape_urls',\n",
    "                     'launched_at', 'name', 'pledged', 'slug', 'state', 'category_type', 'country_name', 'id_x', 'url_x',\n",
    "                       'comments_x', 'dates', 'staff_pick_y', 'Unnamed: 0', 'id_y', 'url_y', 'comments_y', 'have_comments_y', \n",
    "                          'have_comments_x'], axis=1)\n",
    "y1_old = model_data1['state']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_filtered2 = df[['name', 'have_comments', 'pre_mid_campaign_dates_length']].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_data_2 = pd.merge(full, df_filtered2, on='name', how='left')\n",
    "model_data_2.fillna(0,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_data2 = pd.merge(data, model_data_2, on='name', how='right')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_data2.fillna(0, inplace=True)\n",
    "model_data2['state'] = model_data2['state'].replace(to_replace=['successful', 'failed'], value=[1,0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "X2 = model_data2.drop(['backers_count','blurb', 'category', 'deadline', 'urls', 'scrape_url', 'complete_url', 'scrape_urls',\n",
    "                     'launched_at', 'name', 'pledged', 'slug', 'state', 'category_type', 'country_name', 'id_x', 'url_x',\n",
    "                       'comments_x', 'dates', 'staff_pick_y', 'Unnamed: 0', 'id_y', 'url_y', 'comments_y', 'have_comments_y'],\n",
    "                   axis=1)\n",
    "y2 = model_data2['state']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Old model with added \"have comments\" features. FULL DATASET\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model, Accuracy, Precision, Recall\n",
      "    Random Forest: (0.7358033973169185, 0.7582875618402088, 0.7943360809597649)\n",
      "    Logistic Regression: (0.6240161642741349, 0.619026288529914, 0.9022688321227454)\n",
      "    Gradient Boost: (0.7401029062801024, 0.7455435417884279, 0.8328572594466661)\n"
     ]
    }
   ],
   "source": [
    "def get_scores(classifier, X_train, X_test, y_train, y_test, **kwargs):\n",
    "    model = classifier(**kwargs)\n",
    "    model.fit(X_train, y_train)\n",
    "    y_predict = model.predict(X_test)\n",
    "    return model.score(X_test, y_test), precision_score(y_test, y_predict), recall_score(y_test, y_predict)\n",
    "print(\"Model, Accuracy, Precision, Recall\")\n",
    "print(\"    Random Forest:\", get_scores(RandomForestClassifier, X_train, X_test, y_train, y_test))\n",
    "print(\"    Logistic Regression:\", get_scores(LogisticRegression, X_train, X_test, y_train, y_test))\n",
    "print(\"    Gradient Boost:\", get_scores(GradientBoostingClassifier, X_train, X_test, y_train, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Old model with scraped campaign, no added new comment features.\n",
    "X_train, X_test, y_train, y_test = train_test_split(X1_old, y1_old)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model, Accuracy, Precision, Recall\n",
      "    Random Forest: (0.810455563853622, 0.8312179915386328, 0.8795947219604147)\n",
      "    Logistic Regression: (0.6725914861837192, 0.6726691349713901, 0.9418001885014138)\n",
      "    Gradient Boost: (0.8058252427184466, 0.8205574912891986, 0.8878416588124411)\n"
     ]
    }
   ],
   "source": [
    "def get_scores(classifier, X_train, X_test, y_train, y_test, **kwargs):\n",
    "    model = classifier(**kwargs)\n",
    "    model.fit(X_train, y_train)\n",
    "    y_predict = model.predict(X_test)\n",
    "    return model.score(X_test, y_test), precision_score(y_test, y_predict), recall_score(y_test, y_predict)\n",
    "print(\"Model, Accuracy, Precision, Recall\")\n",
    "print(\"    Random Forest:\", get_scores(RandomForestClassifier, X_train, X_test, y_train, y_test))\n",
    "print(\"    Logistic Regression:\", get_scores(LogisticRegression, X_train, X_test, y_train, y_test))\n",
    "print(\"    Gradient Boost:\", get_scores(GradientBoostingClassifier, X_train, X_test, y_train, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Old model with added \"have comments\" features. 26.8K DATAPOINTS, THOSE THAT ARE SCRAPED\n",
    "X_train, X_test, y_train, y_test = train_test_split(X1, y1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model, Accuracy, Precision, Recall\n",
      "    Random Forest: (0.815982076176251, 0.8376030296279795, 0.8820079756040348)\n",
      "    Logistic Regression: (0.6642270351008215, 0.6714310022120129, 0.9256392212057236)\n",
      "    Gradient Boost: (0.8150858849887976, 0.8321985504063255, 0.88881069669247)\n"
     ]
    }
   ],
   "source": [
    "def get_scores(classifier, X_train, X_test, y_train, y_test, **kwargs):\n",
    "    model = classifier(**kwargs)\n",
    "    model.fit(X_train, y_train)\n",
    "    y_predict = model.predict(X_test)\n",
    "    return model.score(X_test, y_test), precision_score(y_test, y_predict), recall_score(y_test, y_predict)\n",
    "print(\"Model, Accuracy, Precision, Recall\")\n",
    "print(\"    Random Forest:\", get_scores(RandomForestClassifier, X_train, X_test, y_train, y_test))\n",
    "print(\"    Logistic Regression:\", get_scores(LogisticRegression, X_train, X_test, y_train, y_test))\n",
    "print(\"    Gradient Boost:\", get_scores(GradientBoostingClassifier, X_train, X_test, y_train, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Old model with added \"have comments\" and \"pre_mid_campaign_dates_length\" features. 26.8K DATAPOINTS, THOSE THAT ARE SCRAPED\n",
    "X_train, X_test, y_train, y_test = train_test_split(X2, y2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model, Accuracy, Precision, Recall\n",
      "    Random Forest: (0.8150858849887976, 0.8403607666290868, 0.8754991778247592)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/Helen/anaconda3/lib/python3.7/site-packages/sklearn/linear_model/_logistic.py:940: ConvergenceWarning: lbfgs failed to converge (status=1):\n",
      "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n",
      "\n",
      "Increase the number of iterations (max_iter) or scale the data as shown in:\n",
      "    https://scikit-learn.org/stable/modules/preprocessing.html\n",
      "Please also refer to the documentation for alternative solver options:\n",
      "    https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n",
      "  extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    Logistic Regression: (0.6788648244958925, 0.6831218494698418, 0.923185341789993)\n",
      "    Gradient Boost: (0.8117998506348021, 0.8311602209944752, 0.8834860230209067)\n"
     ]
    }
   ],
   "source": [
    "def get_scores(classifier, X_train, X_test, y_train, y_test, **kwargs):\n",
    "    model = classifier(**kwargs)\n",
    "    model.fit(X_train, y_train)\n",
    "    y_predict = model.predict(X_test)\n",
    "    return model.score(X_test, y_test), precision_score(y_test, y_predict), recall_score(y_test, y_predict)\n",
    "print(\"Model, Accuracy, Precision, Recall\")\n",
    "print(\"    Random Forest:\", get_scores(RandomForestClassifier, X_train, X_test, y_train, y_test))\n",
    "print(\"    Logistic Regression:\", get_scores(LogisticRegression, X_train, X_test, y_train, y_test))\n",
    "print(\"    Gradient Boost:\", get_scores(GradientBoostingClassifier, X_train, X_test, y_train, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_lstm = df[df['pre_mid_campaign_comments'].apply(lambda x: len(x)>2)]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "def string_convert(x):\n",
    "    str1 = ''.join(str(e) for e in x)\n",
    "    return str1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    }
   ],
   "source": [
    "df_lstm['pre_mid_campaign_comments'] = df_lstm['pre_mid_campaign_comments'].apply(lambda x: string_convert(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = df_lstm['pre_mid_campaign_comments'].values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model with 150 keywords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# tokenizer = Tokenizer(num_words=150)\n",
    "# maxlen=50\n",
    "# embedding_dim=100\n",
    "# model=Sequential()\n",
    "# model.add(layers.Embedding(input_dim=vocab_size,\n",
    "#          output_dim=embedding_dim,\n",
    "#          input_length=maxlen))\n",
    "# model.add(layers.LSTM(units=50,return_sequences=True))\n",
    "# model.add(layers.LSTM(units=30))\n",
    "# model.add(layers.Dropout(0.2))\n",
    "# model.add(layers.Dense(1, activation=\"sigmoid\"))\n",
    "# model.compile(optimizer=\"adam\", loss=\"binary_crossentropy\", \n",
    "#      metrics=['accuracy'])\n",
    "# model.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = Tokenizer(num_words=150)\n",
    "tokenizer.fit_on_texts(x)\n",
    "xtest= tokenizer.texts_to_sequences(x) \n",
    "maxlen=50\n",
    "xtest=pad_sequences(xtest,padding='post', maxlen=maxlen)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 38  63  18  82   6  20  75   8  95  17   4  45  16 103  16   5  17   2\n",
      "  16  70   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0   0   0   0   0]\n",
      "Lori, the kids will want you and Uncle Michael to sign their card!  Enjoy your trip.My $25 pledge was on behalf of Nina Gibbons, who donated to support this project.  Thanks so much for your donation.Hope you guys have a great trip! My only request-my brother and Lori have to sign my book!\n"
     ]
    }
   ],
   "source": [
    "print(xtest[0])\n",
    "print(x[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "24495"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vocab_size=len(tokenizer.word_index)+1\n",
    "vocab_size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_150 = keras.models.load_model('my_half_model_150')\n",
    "predict_150 = model_150.predict(xtest)\n",
    "df_lstm['sentiment_predict'] = predict_150"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Model with 200 Keywords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# tokenizer = Tokenizer(num_words=200)\n",
    "# maxlen=80\n",
    "# embedding_dim=100\n",
    "# model=Sequential()\n",
    "# model.add(layers.Embedding(input_dim=vocab_size,\n",
    "#          output_dim=embedding_dim,\n",
    "#          input_length=maxlen))\n",
    "# model.add(layers.LSTM(units=50,return_sequences=True))\n",
    "# model.add(layers.LSTM(units=20))\n",
    "# model.add(layers.Dropout(0.2))\n",
    "# model.add(layers.Dense(1, activation=\"sigmoid\"))\n",
    "# model.compile(optimizer=\"adam\", loss=\"binary_crossentropy\", \n",
    "#      metrics=['accuracy'])\n",
    "# model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = Tokenizer(num_words=200)\n",
    "tokenizer.fit_on_texts(x)\n",
    "xtest= tokenizer.texts_to_sequences(x) \n",
    "maxlen=80\n",
    "xtest=pad_sequences(xtest,padding='post', maxlen=maxlen)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[  1  21  88   8   5   2 173  20  16  44  48  14   7 162   2 119   9  38\n",
      "  63  18  82   6  20  75   8  95  17   4  45  16 103  16   5  17   2  16\n",
      "  70   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0]\n",
      "Lori, the kids will want you and Uncle Michael to sign their card!  Enjoy your trip.My $25 pledge was on behalf of Nina Gibbons, who donated to support this project.  Thanks so much for your donation.Hope you guys have a great trip! My only request-my brother and Lori have to sign my book!\n"
     ]
    }
   ],
   "source": [
    "print(xtest[0])\n",
    "print(x[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "24495"
      ]
     },
     "execution_count": 200,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vocab_size=len(tokenizer.word_index)+1\n",
    "vocab_size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_200 = keras.models.load_model('my_model_200')\n",
    "predict_200 = model_200.predict(xtest)\n",
    "df_lstm['sentiment_predict'] = predict_200"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    }
   ],
   "source": [
    "df_lstm['sentiment_predict'] = df_lstm['sentiment_predict'].map(lambda x:1 if (x > 0.5) else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LSTM OHE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "lstm_filtered = df_lstm[['name', 'sentiment_predict']].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "lstm_data = pd.merge(full, lstm_filtered, on='name', how='left')\n",
    "lstm_data.fillna(0,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "lstm_data1 = pd.merge(model_data1, lstm_data, on='name', how='right')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0    19660\n",
      "1.0     9265\n",
      "Name: have_comments_y, dtype: int64\n",
      "(28925, 62)\n"
     ]
    }
   ],
   "source": [
    "print(lstm_data1['have_comments_y'].value_counts())\n",
    "print(lstm_data1.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "lstm_data1.fillna(0, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "lstm_x = lstm_data1.drop(['backers_count','blurb', 'category', 'deadline', 'name', 'state',\n",
    "                          'urls', 'id_x', 'scrape_url', 'complete_url', 'scrape_urls', 'url_x', 'id_y',\n",
    "                     'launched_at', 'pledged', 'slug', 'state', 'category_type', 'country_name',\n",
    "                          'url_y', 'comments_y', 'comments_x', 'dates', 'staff_pick_y', 'id', 'url',\n",
    "                         'comments', 'Unnamed: 0_y', 'Unnamed: 0_x', 'have_comments_y'],\n",
    "                   axis=1)\n",
    "lstm_y = lstm_data1['state']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "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>staff_pick_x</th>\n",
       "      <th>launched_year</th>\n",
       "      <th>launched_month</th>\n",
       "      <th>campaign_length</th>\n",
       "      <th>blurb_length</th>\n",
       "      <th>have_comments_x</th>\n",
       "      <th>category_type_art</th>\n",
       "      <th>category_type_comics</th>\n",
       "      <th>category_type_crafts</th>\n",
       "      <th>category_type_dance</th>\n",
       "      <th>category_type_design</th>\n",
       "      <th>category_type_fashion</th>\n",
       "      <th>category_type_film &amp; video</th>\n",
       "      <th>category_type_food</th>\n",
       "      <th>category_type_games</th>\n",
       "      <th>category_type_journalism</th>\n",
       "      <th>category_type_music</th>\n",
       "      <th>category_type_photography</th>\n",
       "      <th>category_type_publishing</th>\n",
       "      <th>category_type_technology</th>\n",
       "      <th>category_type_theater</th>\n",
       "      <th>country_name_Australia</th>\n",
       "      <th>country_name_Canada</th>\n",
       "      <th>country_name_France</th>\n",
       "      <th>country_name_Germany</th>\n",
       "      <th>country_name_Italy</th>\n",
       "      <th>country_name_Mexico</th>\n",
       "      <th>country_name_Spain</th>\n",
       "      <th>country_name_other</th>\n",
       "      <th>country_name_the Netherlands</th>\n",
       "      <th>country_name_the United States</th>\n",
       "      <th>sentiment_predict</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1000.0</td>\n",
       "      <td>False</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1100.0</td>\n",
       "      <td>False</td>\n",
       "      <td>2019.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>25000.0</td>\n",
       "      <td>False</td>\n",
       "      <td>2016.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>25.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>500.0</td>\n",
       "      <td>False</td>\n",
       "      <td>2012.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2500.0</td>\n",
       "      <td>False</td>\n",
       "      <td>2017.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>31.0</td>\n",
       "      <td>23.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28920</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28921</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28922</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28923</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28924</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>28925 rows × 33 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          goal staff_pick_x  launched_year  launched_month  campaign_length  \\\n",
       "0       1000.0        False         2017.0             9.0             14.0   \n",
       "1       1100.0        False         2019.0             7.0             60.0   \n",
       "2      25000.0        False         2016.0             1.0             60.0   \n",
       "3        500.0        False         2012.0             9.0             10.0   \n",
       "4       2500.0        False         2017.0            11.0             31.0   \n",
       "...        ...          ...            ...             ...              ...   \n",
       "28920      0.0            0            0.0             0.0              0.0   \n",
       "28921      0.0            0            0.0             0.0              0.0   \n",
       "28922      0.0            0            0.0             0.0              0.0   \n",
       "28923      0.0            0            0.0             0.0              0.0   \n",
       "28924      0.0            0            0.0             0.0              0.0   \n",
       "\n",
       "       blurb_length  have_comments_x  category_type_art  category_type_comics  \\\n",
       "0              24.0              0.0                0.0                   0.0   \n",
       "1              15.0              0.0                0.0                   0.0   \n",
       "2              25.0              0.0                0.0                   0.0   \n",
       "3              12.0              0.0                0.0                   0.0   \n",
       "4              23.0              0.0                0.0                   0.0   \n",
       "...             ...              ...                ...                   ...   \n",
       "28920           0.0              0.0                0.0                   0.0   \n",
       "28921           0.0              0.0                0.0                   0.0   \n",
       "28922           0.0              0.0                0.0                   0.0   \n",
       "28923           0.0              0.0                0.0                   0.0   \n",
       "28924           0.0              0.0                0.0                   0.0   \n",
       "\n",
       "       category_type_crafts  category_type_dance  category_type_design  \\\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                       0.0                  0.0                   0.0   \n",
       "...                     ...                  ...                   ...   \n",
       "28920                   0.0                  0.0                   0.0   \n",
       "28921                   0.0                  0.0                   0.0   \n",
       "28922                   0.0                  0.0                   0.0   \n",
       "28923                   0.0                  0.0                   0.0   \n",
       "28924                   0.0                  0.0                   0.0   \n",
       "\n",
       "       category_type_fashion  category_type_film & video  category_type_food  \\\n",
       "0                        0.0                         0.0                 0.0   \n",
       "1                        0.0                         1.0                 0.0   \n",
       "2                        0.0                         0.0                 0.0   \n",
       "3                        0.0                         0.0                 0.0   \n",
       "4                        0.0                         0.0                 0.0   \n",
       "...                      ...                         ...                 ...   \n",
       "28920                    0.0                         0.0                 0.0   \n",
       "28921                    0.0                         0.0                 0.0   \n",
       "28922                    0.0                         0.0                 0.0   \n",
       "28923                    0.0                         0.0                 0.0   \n",
       "28924                    0.0                         0.0                 0.0   \n",
       "\n",
       "       category_type_games  category_type_journalism  category_type_music  \\\n",
       "0                      1.0                       0.0                  0.0   \n",
       "1                      0.0                       0.0                  0.0   \n",
       "2                      1.0                       0.0                  0.0   \n",
       "3                      0.0                       0.0                  0.0   \n",
       "4                      0.0                       0.0                  1.0   \n",
       "...                    ...                       ...                  ...   \n",
       "28920                  0.0                       0.0                  0.0   \n",
       "28921                  0.0                       0.0                  0.0   \n",
       "28922                  0.0                       0.0                  0.0   \n",
       "28923                  0.0                       0.0                  0.0   \n",
       "28924                  0.0                       0.0                  0.0   \n",
       "\n",
       "       category_type_photography  category_type_publishing  \\\n",
       "0                            0.0                       0.0   \n",
       "1                            0.0                       0.0   \n",
       "2                            0.0                       0.0   \n",
       "3                            0.0                       1.0   \n",
       "4                            0.0                       0.0   \n",
       "...                          ...                       ...   \n",
       "28920                        0.0                       0.0   \n",
       "28921                        0.0                       0.0   \n",
       "28922                        0.0                       0.0   \n",
       "28923                        0.0                       0.0   \n",
       "28924                        0.0                       0.0   \n",
       "\n",
       "       category_type_technology  category_type_theater  \\\n",
       "0                           0.0                    0.0   \n",
       "1                           0.0                    0.0   \n",
       "2                           0.0                    0.0   \n",
       "3                           0.0                    0.0   \n",
       "4                           0.0                    0.0   \n",
       "...                         ...                    ...   \n",
       "28920                       0.0                    0.0   \n",
       "28921                       0.0                    0.0   \n",
       "28922                       0.0                    0.0   \n",
       "28923                       0.0                    0.0   \n",
       "28924                       0.0                    0.0   \n",
       "\n",
       "       country_name_Australia  country_name_Canada  country_name_France  \\\n",
       "0                         0.0                  1.0                  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                         0.0                  0.0                  0.0   \n",
       "...                       ...                  ...                  ...   \n",
       "28920                     0.0                  0.0                  0.0   \n",
       "28921                     0.0                  0.0                  0.0   \n",
       "28922                     0.0                  0.0                  0.0   \n",
       "28923                     0.0                  0.0                  0.0   \n",
       "28924                     0.0                  0.0                  0.0   \n",
       "\n",
       "       country_name_Germany  country_name_Italy  country_name_Mexico  \\\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                       0.0                 0.0                  0.0   \n",
       "...                     ...                 ...                  ...   \n",
       "28920                   0.0                 0.0                  0.0   \n",
       "28921                   0.0                 0.0                  0.0   \n",
       "28922                   0.0                 0.0                  0.0   \n",
       "28923                   0.0                 0.0                  0.0   \n",
       "28924                   0.0                 0.0                  0.0   \n",
       "\n",
       "       country_name_Spain  country_name_other  country_name_the Netherlands  \\\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                     0.0                 0.0                           0.0   \n",
       "...                   ...                 ...                           ...   \n",
       "28920                 0.0                 0.0                           0.0   \n",
       "28921                 0.0                 0.0                           0.0   \n",
       "28922                 0.0                 0.0                           0.0   \n",
       "28923                 0.0                 0.0                           0.0   \n",
       "28924                 0.0                 0.0                           0.0   \n",
       "\n",
       "       country_name_the United States  sentiment_predict  \n",
       "0                                 0.0                0.0  \n",
       "1                                 1.0                0.0  \n",
       "2                                 1.0                0.0  \n",
       "3                                 1.0                0.0  \n",
       "4                                 1.0                0.0  \n",
       "...                               ...                ...  \n",
       "28920                             0.0                0.0  \n",
       "28921                             0.0                0.0  \n",
       "28922                             0.0                0.0  \n",
       "28923                             0.0                0.0  \n",
       "28924                             0.0                0.0  \n",
       "\n",
       "[28925 rows x 33 columns]"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lstm_x"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LSTM Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [],
   "source": [
    "## model_150\n",
    "X_train, X_test, y_train, y_test = train_test_split(lstm_x, lstm_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model, Accuracy, Precision, Recall\n",
      "    Random Forest: (0.8344856194690266, 0.8606694560669456, 0.8856835306781485)\n",
      "    Logistic Regression: (0.667450221238938, 0.6622718052738337, 0.9840688912809472)\n",
      "    Gradient Boost: (0.8172013274336283, 0.8346424974823766, 0.8921420882669537)\n"
     ]
    }
   ],
   "source": [
    "def get_scores(classifier, X_train, X_test, y_train, y_test, **kwargs):\n",
    "    model = classifier(**kwargs)\n",
    "    model.fit(X_train, y_train)\n",
    "    y_predict = model.predict(X_test)\n",
    "    return model.score(X_test, y_test), precision_score(y_test, y_predict), recall_score(y_test, y_predict)\n",
    "print(\"Model, Accuracy, Precision, Recall\")\n",
    "print(\"    Random Forest:\", get_scores(RandomForestClassifier, X_train, X_test, y_train, y_test))\n",
    "print(\"    Logistic Regression:\", get_scores(LogisticRegression, X_train, X_test, y_train, y_test))\n",
    "print(\"    Gradient Boost:\", get_scores(GradientBoostingClassifier, X_train, X_test, y_train, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model_200\n",
    "X_train, X_test, y_train, y_test = train_test_split(lstm_x, lstm_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model, Accuracy, Precision, Recall\n",
      "    Random Forest: (0.840016592920354, 0.863267429760666, 0.8924268502581756)\n",
      "    Logistic Regression: (0.6660674778761062, 0.6619759175975627, 0.9817125645438899)\n",
      "    Gradient Boost: (0.8212112831858407, 0.8384103288279201, 0.8941480206540448)\n"
     ]
    }
   ],
   "source": [
    "def get_scores(classifier, X_train, X_test, y_train, y_test, **kwargs):\n",
    "    model = classifier(**kwargs)\n",
    "    model.fit(X_train, y_train)\n",
    "    y_predict = model.predict(X_test)\n",
    "    return model.score(X_test, y_test), precision_score(y_test, y_predict), recall_score(y_test, y_predict)\n",
    "print(\"Model, Accuracy, Precision, Recall\")\n",
    "print(\"    Random Forest:\", get_scores(RandomForestClassifier, X_train, X_test, y_train, y_test))\n",
    "print(\"    Logistic Regression:\", get_scores(LogisticRegression, X_train, X_test, y_train, y_test))\n",
    "print(\"    Gradient Boost:\", get_scores(GradientBoostingClassifier, X_train, X_test, y_train, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [],
   "source": [
    "# With sentiment predict cast as 1 and 0\n",
    "X_train, X_test, y_train, y_test = train_test_split(lstm_x, lstm_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model, Accuracy, Precision, Recall\n",
      "    Random Forest: (0.8357300884955752, 0.8516844700082169, 0.898959236773634)\n",
      "    Logistic Regression: (0.6527931415929203, 0.6559762435040831, 0.9579358196010408)\n",
      "    Gradient Boost: (0.8160951327433629, 0.828594313175811, 0.8972246313963573)\n"
     ]
    }
   ],
   "source": [
    "def get_scores(classifier, X_train, X_test, y_train, y_test, **kwargs):\n",
    "    model = classifier(**kwargs)\n",
    "    model.fit(X_train, y_train)\n",
    "    y_predict = model.predict(X_test)\n",
    "    return model.score(X_test, y_test), precision_score(y_test, y_predict), recall_score(y_test, y_predict)\n",
    "print(\"Model, Accuracy, Precision, Recall\")\n",
    "print(\"    Random Forest:\", get_scores(RandomForestClassifier, X_train, X_test, y_train, y_test))\n",
    "print(\"    Logistic Regression:\", get_scores(LogisticRegression, X_train, X_test, y_train, y_test))\n",
    "print(\"    Gradient Boost:\", get_scores(GradientBoostingClassifier, X_train, X_test, y_train, y_test))"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}