Board_Game_Review_Predictions_assignment.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Board game review prediction\n",
    "We will implement Linear Regression and Random Forest Regressor from sklearn to predict average rating of a board game. The <a href=\"https://github.com/ThaWeatherman/scrapers/blob/master/boardgamegeek/games.csv\">data</a> is downloaded from a github profile by TheWeatherman."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing libraries and functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn import model_selection\n",
    "from sklearn.metrics import mean_squared_error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "games = pd.read_csv('games.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index([u'id', u'type', u'name', u'yearpublished', u'minplayers', u'maxplayers',\n",
      "       u'playingtime', u'minplaytime', u'maxplaytime', u'minage',\n",
      "       u'users_rated', u'average_rating', u'bayes_average_rating',\n",
      "       u'total_owners', u'total_traders', u'total_wanters', u'total_wishers',\n",
      "       u'total_comments', u'total_weights', u'average_weight'],\n",
      "      dtype='object')\n",
      "(81312, 20)\n"
     ]
    }
   ],
   "source": [
    "print(games.columns)\n",
    "print(games.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EDA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Histogram of average_rating"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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": [
    "plt.hist(games['average_rating'])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "id                             318\n",
      "type                     boardgame\n",
      "name                    Looney Leo\n",
      "yearpublished                    0\n",
      "minplayers                       0\n",
      "maxplayers                       0\n",
      "playingtime                      0\n",
      "minplaytime                      0\n",
      "maxplaytime                      0\n",
      "minage                           0\n",
      "users_rated                      0\n",
      "average_rating                   0\n",
      "bayes_average_rating             0\n",
      "total_owners                     0\n",
      "total_traders                    0\n",
      "total_wanters                    0\n",
      "total_wishers                    1\n",
      "total_comments                   0\n",
      "total_weights                    0\n",
      "average_weight                   0\n",
      "Name: 13048, dtype: object\n",
      "id                                  12333\n",
      "type                            boardgame\n",
      "name                    Twilight Struggle\n",
      "yearpublished                        2005\n",
      "minplayers                              2\n",
      "maxplayers                              2\n",
      "playingtime                           180\n",
      "minplaytime                           180\n",
      "maxplaytime                           180\n",
      "minage                                 13\n",
      "users_rated                         20113\n",
      "average_rating                    8.33774\n",
      "bayes_average_rating              8.22186\n",
      "total_owners                        26647\n",
      "total_traders                         372\n",
      "total_wanters                        1219\n",
      "total_wishers                        5865\n",
      "total_comments                       5347\n",
      "total_weights                        2562\n",
      "average_weight                     3.4785\n",
      "Name: 0, dtype: object\n"
     ]
    }
   ],
   "source": [
    "print(games[games['average_rating'] == 0].iloc[0])\n",
    "print(games[games['average_rating'] > 0].iloc[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Removing entries with users_rated = 0 and missing data\n",
    "games = games[games[\"users_rated\"] > 0]\n",
    "games = games.dropna()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Let's visualize the histogram of average_rating after the cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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": [
    "plt.hist(games[\"average_rating\"])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Correlation matrix across features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x648 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "corr_mat = games.corr()\n",
    "plt.figure(figsize=(12, 9))\n",
    "sns.heatmap(corr_mat, vmax=0.8, square=True)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Filtering columns to remove unneccessary features from the data\n",
    "columns = games.columns\n",
    "columns = [c for c in columns if c not in [\"bayes_average_rating\", \"average_rating\",\n",
    "                                           \"id\", \"type\", \"name\", \"yearpublished\"]]\n",
    "# Variable to predict\n",
    "target = \"average_rating\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Linear Regression and Random Forest Regression "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(45515, 20)\n",
      "(11379, 20)\n"
     ]
    }
   ],
   "source": [
    "# Generating training set using random sampling\n",
    "train = games.sample(frac=0.8, random_state=1)\n",
    "test = games[~games.index.isin(train.index)]\n",
    "print(train.shape)\n",
    "print(test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.088345766529933\n"
     ]
    }
   ],
   "source": [
    "lin_reg = LinearRegression()\n",
    "lin_reg.fit(train[columns], train[target])\n",
    "lr_preds = lin_reg.predict(test[columns])\n",
    "print(mean_squared_error(test[target], lr_preds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.5689077717887692\n"
     ]
    }
   ],
   "source": [
    "rfr_reg = RandomForestRegressor(n_estimators=100, min_samples_leaf=10, random_state=1)\n",
    "rfr_reg.fit(train[columns], train[target])\n",
    "rfr_preds = rfr_reg.predict(test[columns])\n",
    "print(mean_squared_error(test[target], rfr_preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Using k-fold for evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear Regression using k-fold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean_score: 0.15320640585954814\n"
     ]
    }
   ],
   "source": [
    "kfold = model_selection.KFold(n_splits=10, shuffle = True)\n",
    "lin_cv_res = model_selection.cross_val_score(lin_reg, train[columns], train[target], cv=kfold, scoring='r2')\n",
    "print(\"mean_score: \" + str(lin_cv_res.mean()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest Regression using k-fold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mean_score: 0.3668440895645602\n"
     ]
    }
   ],
   "source": [
    "rfr_cv_res = model_selection.cross_val_score(rfr_reg, train[columns], train[target], cv=kfold, scoring='r2')\n",
    "print(\"mean_score: \" + str(rfr_cv_res.mean()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}