Data Sampling.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rc('font', family = \"serif\")\n",
    "plt.rc('xtick', labelsize='x-small')\n",
    "plt.rc('ytick', labelsize='x-small')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading the data :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"data/ownedgames.json\", \"r\") as f :\n",
    "    data = json.loads(f.read())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Removing null entries :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = {key : value for key, value in data.items() if value != {}}\n",
    "data = {key : value for key, value in data.items() if value[\"game_count\"] != 0}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Making a DataFrame :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_values = []\n",
    "for key, value in data.items() :\n",
    "    steamid = key\n",
    "    for item in value[\"games\"] :\n",
    "        df_values.append((int(steamid), item[\"appid\"], item[\"playtime_forever\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame(df_values, columns = [\"steamid\", \"appid\", \"playtime_forever\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>steamid</th>\n",
       "      <th>appid</th>\n",
       "      <th>playtime_forever</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>30</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>50</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             steamid  appid  playtime_forever\n",
       "0  76561197969918174     10                 3\n",
       "1  76561197969918174     20                 0\n",
       "2  76561197969918174     30                 0\n",
       "3  76561197969918174     40                 0\n",
       "4  76561197969918174     50                 0"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Removing Zero Playtime Data :\n",
    "We remove the zero-playtime data for building models since we treat zero-playtime and not playing as equal. This renders smaller sized data without losing any information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df[df[\"playtime_forever\"] != 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>steamid</th>\n",
       "      <th>appid</th>\n",
       "      <th>playtime_forever</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>70</td>\n",
       "      <td>57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>80</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>220</td>\n",
       "      <td>5685</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>240</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              steamid  appid  playtime_forever\n",
       "0   76561197969918174     10                 3\n",
       "6   76561197969918174     70                57\n",
       "7   76561197969918174     80                 3\n",
       "10  76561197969918174    220              5685\n",
       "11  76561197969918174    240                35"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Density of raw data :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "density = lambda df : len(df) / (df[\"steamid\"].nunique() * df[\"appid\"].nunique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unique SteamIds : 15582\n",
      "Unique AppIds : 21925\n",
      "Density : 0.99 %\n"
     ]
    }
   ],
   "source": [
    "print(\"Unique SteamIds :\", df[\"steamid\"].nunique())\n",
    "print(\"Unique AppIds :\", df[\"appid\"].nunique())\n",
    "print(\"Density : {0:.2f} %\".format(density(df) * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sampling :\n",
    "Since games that have few players are not useful for recommendation, we discard such games. Similarly, players that have few games in their library are not useful for recommendation and we discard such players as well. The two criterias based on which we discard are \n",
    "1. Discard a `appid` if it has less than `c1` unique players\n",
    "2. Discard a `steamid` if it has less than `c2` unique games played\n",
    "\n",
    "We determine cutoffs `c1` and `c2` by the distribution of `steamid` counts and `appid` counts respectively. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Grouping by `appid` to get player count for each `appid` :"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "groupby_appid = df.groupby(by = \"appid\")[\"steamid\"].count()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### To determine `c1`, we plot the number of `appids` against unique `steamids`. \n",
    "For example, (1, 100) corresponds to there being 100 games that have only 1 player."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "elbow_appid = groupby_appid.groupby(groupby_appid).count()\n",
    "x, y = elbow_appid.index, elbow_appid.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "appid\n",
       "10    3942\n",
       "20    1407\n",
       "30    1111\n",
       "40     713\n",
       "50    1313\n",
       "Name: steamid, dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "groupby_appid.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2371"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(groupby_appid[groupby_appid == 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "steamid\n",
       "1    2371\n",
       "2    1354\n",
       "3     876\n",
       "4     861\n",
       "5     742\n",
       "Name: steamid, dtype: int64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "elbow_appid.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARoAAADQCAYAAADYt67cAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAHIJJREFUeJzt3XmcFdWZ//HPt4FmERoRWhRkMyphxA1RYwyIcUlM1IRMok5coln4aX4uYyZxmbhAjFEZoyYatzGOJlF/idEY45IMPxHjligoLpHgxiJosBFZbFlafOaPp+70temluuniVsnzfr3u61bdrqr73Av13HNOnXNKZkYIIWSpqtIBhBA++iLRhBAyF4kmhJC5SDQhhMxFogkhZC4STQghc5FoQgiZi0QTQshcJJoQQua6VjqAjhgwYIANHz680mGEsNmZNWvWUjOrbe9+hUw0w4cPZ+bMmZUOI4TNjqQFHdkvqk4hhMxFogkhZC4STQghc5FoQgiZK2yieflluPJKWLq00pGEENpS2ESzbBk8+CCsWFHpSEIIbSlsoqmu9ud16yobRwihbYVPNGvXVjaOEELbCp9ookQTQv4VNtF07+7PkWhCyL/CJpqoOoVQHIVPNFGiCSH/CptoouoUQnEUNtFE1SmE4ihsounaFaQo0YRQBIVNNJKXaiLRhJB/hU004Ikmqk4h5F+hE0337lGiCaEICp1oouoUQjEUPtFE1SmE/Ct8ookSTQj5F4kmhJC5QieaaAwOoRgKnWiijSaEYih8ookSTQj5F4kmhJC5Qiea7t2j6hRCEWRy721Jnwd2AboALyUvDwQGA1PwBHc+sAh408zulPSV8m3MbE1b7xMlmhCKIZNEAzxjZvdJ6gv8HOhmZl9IkslEPNHMMrM7JP0OuBM4tsk2t5cfUNIkYBLA0KFDAU80DQ1g5oMsQwj5lEnVyczeSBa/CFwG9EjW64BhwNBkGaBX8tx0m6bHvMHMxprZ2NraWiAmvwqhKLIq0SDpM8A8vHpUqgbVAgvxBFebvLY6eW66TZvKp/MsJZ0QQv5k1UZzKN4G8yzQF7hN0inAEGAyIGCypIHArcluTbdpU/kse336dF78IYTOlUmiMbMHgAfa2OzMJvv8ur3vE1WnEIqh0Je3404IIRTDRyLRRF+aEPKt0Ikmqk4hFEObiUZSX0k9JXWX9EVJW26KwNKIqlMIxZCmRHMd8DFgKnA0cHmmEbVDVJ1CKIY0iWYW8DdgXzM7msYhBRUXVacQiiFNovk48O/AY8l6bSvbblJRdQqhGNIkmpuAAcBFkg7De/vmQlSdQiiGNjvsmdnjkl4CtgFmmNm72YeVTlSdQiiGNFedjgUewYcUHCjp+5lHlVJUnUIohjRVp93MbBQ+rcPvaRxtXXFVVdC1aySaEPIuTaJZkTxb8pyrFpGYoDyE/EszqHKgpOuAbSVdRmPCyYWYZS+E/EtTovlX4GngdWAucFamEbVTJJoQ8i/NVaf1wA2ldUkHAg9mGVR7xE3kQsi/NhNNMonVyUBvfMKqofiQhFyIEk0I+ZemjeYc4DTgHTzRfC3TiNopGoNDyL80iWammc0urUj6fYbxtFv37rCmzRuzhBAqKU2ieULSw8CreIlmF2BsplG1Q3U1rFxZ6ShCCK1Jk2jOAC4Flifrx2UXTvtF1SmE/EuTaJ5KegQDIGlxhvG0W1x1CiH/0iSafpJuwatOAOOBg7ILqX3iqlMI+Zcm0YwAbixbX97ShpUQVacQ8i9NojnRzF4prUhq635Nm1RUnULIvzQ9g1+RNIrGmfWOA76VaVTtUF0NH3wA77/vI7lDCPmTpmfwVGAkMAifL3hUin26At8DRpjZpI0NsjXlc9JEogkhn9IMqlxtZl8Afmtmx9B4r+zW9AKmlY4v6SJJNyaP3pJ6SbpE0imS/jnZ5ivJ+sWSeqT9ADHLXgj5lybRJGUG+iUllT3b2sHMVgJLy16qBx4G/gq8B0zEJ9K6Gjg22ebYZP3p5O8fImmSpJmSZtbV1TUGF7PshZB7aRJNg6TDgZnAKqAjp/TPzOyXQAPwZXxgZilblGbsK5Vi6oBhTQ9gZjeY2VgzG1tb23gjhpigPIT8S9OqMbU0IbmkaWa2oq0dmjESeBJ4G08ir9PYuLw6eS6NWKoFFqY9cFSdQsi/NInmp5JuLq1IMmCemS1qY7+jgJGSdgcmSdoZ2A24EE8ukyUNpLHN5zZJpwBDgMlpP0C3bv4ciSaE/EqTaHYCfojfz2l7kpKHpHvM7KqWdjKzS/ExUgDfbGaTM5ts/+s0ATcVVacQ8i9NG80DZjbezL5mZuOA/29mBwMDM44tlag6hZB/aRLNoCbr2yfPuZicIa46hZB/aRLN+5Luk3SlpPuBdZL2JScDK6PqFEL+pRmCcLqkzwE7A9PM7L7kT4dkGllKpapTJJoQ8itVp30zux+4P+NYOmSLLfy5vr6ycYQQWpam6pRr3bt79WnVqkpHEkJoSYuJRtIxkrpsymA6qqYm5g0OIc9aK9HsZmbrJR1d/mLSAS9X+vSJEk0IedZaG00/SROAQyS9UfZ6ruajgUg0IeRda4nmNuCrwB74bVZKdsk0og6oqYF58yodRQihJS0mGjN7CHhI0jgze6T0uqT9Nklk7dCnT7TRhJBnaS5vPyrpm/iAyNnATdmG1H59+sC774IZSG1vH0LYtNJc3r4CGIMPqhybrOdKTY0nmXffrXQkIYTmpCnR1JnZRaUVSZOzC6djamr8edUqL92EEPIlTYmmb5P1miwC2Ril5BJXnkLIpzQlmpclzQbm4zeTa3EOmkopJZpoEA4hn9IMqvxPSY8Ao4HnzWxu9mG1T3nVKYSQP2kHVf4d+HvGsXRYVJ1CyLfCD6oEH8EtRdUphLz6SCQaKTrthZBnbSYaSQskjdkUwWyMmpqoOoWQV2lKNHeb2dOlFUnbt7ZxpcTAyhDyK+2cwSdJ2l/SeOCcrIPqiJiTJoT8SnPV6QjgUWCfZD13o7fBSzSvvFLpKEIIzUmTaM4ws3tLK5L2amsHSV2B7wEjzGzSRsSXWlSdQsivNB327pV0OD70YDbwtxTH7QVMA04CkPQV/IZzg4EpeJXtfGAR8KaZ3dl0GzNb09yBW1JT4/d2Wru28c4IIYR8SHPVaSrwJWA8UA1c0tY+ZrYSWFr20rFmdjXwNDAxecxKXju2hW2axjFJ0kxJM+vq6jZ4z+i0F0J+pWkMXm5mJwKvmdkzwLIOvE+P5LkOGAYMTZbBSz/NbfMhZnaDmY01s7G1tbUbvEGMdwohv9K00QxIni157shEDKVqUC2wEE9wpWyxuoVt2iXGO4WQX2kSzUuSXgQ+kPRl4PqUxz4KGJncNeE2SacAQ4DJ+BzEkyUNBG5Ntm+6TbtEogkhv9I0Bl8naQbtHL1tZpcClyars5vZ5Mwm2/86zXFbElWnEPKrzUQjqRtwAH7v7a0lzTOzdZlH1k6RaELIrzSNwTfhnfReA3Ylh5OTA3TtCj17RtUphDxK00azxMy+W1qR9JMM49koW24JDz/spZtDD4W+TSchDSFURJoSzRtN1t/OIpDOcPrp8LGPwa23wlln+Z0RQgiV12KikfSQpOnAMZLmS5ohaT5w5CaLrp123hmmTIFvfxsWL4b58ysdUQgBWi/RPAmciPcK3h/4GjABuC/7sDbOJz7hz08+Wdk4QgiuxURjZmeZ2YImj/nAHZsuvI7p1w922gmeeqrSkYQQIN3l7TF4aaYP3tFuF/yOlbm2997eVrN8uTcShxAqJ01j8FXAQ8AtwM003/kud/bayxuDo1QTQuWlubw9y8zuLq1IWpxhPJ1mxAgYMMATzcEHVzqaEDZvaRLNQ5JuAV5N1scDB2UXUueQvPo0fXrMURNCpaWpOn0PnyNmQfJYnmlEnWjCBFizBu66q9KRhLB5S1OimW1m/9sbWNJjGcbTqUaNgnHj4Le/hQMPhK23rnREIWye0t4FYYqk4yUdD1ycdVCd6etf9+ebcjlCK4TNQ5pEMw74ABiRPLbKNKJONmAAHHkkPPYYzJpV6WhC2DylqTqdbGZ/Ka1I2jHDeDIxcSI8+ihcfjn85CeefEIIm06bJZryJJM4IKNYMlNdDWef7XdJmDoV3n+/0hGFsHlJcxeEdyS9JmmepHfx26QUzuDBcNppMGeONw6HEDadNG00k8xsezMbgU9UfkHGMWVm3DjvMXzvvdDQUOloQth8pKk63VG2vAZvEC6sww6DFSu8cTiEsGmkGVT5EI23WindrbKw9tgDBg2C++7zDn0hhOyluer0F+C6ZHmVmXXkBnK5IcHnPgc33givvuoz8oUQstXaDHujAczsnLL5aAqdZEoOOsjHPt17b6UjCWHz0FqJ5geS7m7uD2b2i4zi2SS22MKrTTNmwLe+Bb16tbVHCGFjtNYYvAKYT+NgypXAD4FPZx9W9g4+2Ed1P/JIpSMJ4aOvtRLNuWa2GEDSbsANwBVmdkV730TScHwCrSXAdKABGAgMBqbgCe98YBHwppnd2d73aK+ddoIhQ2DaNPjMZ7J+txA2b63NGVxKMscB9+BDEdqdZMrMAWYCzwPHmtnV+PQTE5PHrOS1Y5vbWdIkSTMlzayrq9uIMErH81LN3Lnw+usbfbgQQitaawzuIulq4AxgfzObnrzeswPvsxiYDFwPXAL0SF6vA4YBQ5NlgGZbTMzsBjMba2Zja2trOxDChg44ALp08VJNCCE7rbXR/BmfiPyrwAeShkoahld12mtHoMrMDK+urUlerwUWAq8nywCrO3D8DtlyS+8pPH26d+ILIWSjtTaadcAM/IZxKnt9TAfeZxBwgqSFwF3AckmnAEPwko6AyZIGArd24PgdduSRPuDy3HPhRz/y2+mGEDqXrIX7xkoaZ2YbXJORtJ+ZVbQD/9ixY23mzJmddrzZs+EHP4Dhw+GHP4zL3SG0RNIsM2v37ZZaawxu9sJvpZNMFnbfHc45x3sKX3ll3LM7hM6WZvT2ZmGvvXzazyeegDtyfy/OEIolEk2ZI46A/feHX/0K/tJ0uq8QQodFoikjwamnwg47wMUXw0MPVTqiED4aItE00b07XHQRjB7tcwzff3+lIwqh+CLRNKNnT7jgAr/T5XXXweOPVzqiEIotEk0LqqvhrLN8TNSPf+xDFUIIHROJphXV1XDeebDVVjBlCrzwQqUjCqGYItG0oW9fuPBCfz7vPG+zee21GIgZQnukmcpzs7fNNnDZZXDJJXDttY2vH3ecD2EIIbQuEk1KW2wBkyfDc8/5jegeeMA79h1yiA/ODCG0LKpO7dCli99FYZ99YNIkTzi3317pqELIv0g0HTRoEHz2s/CnP8EbbzS+/sEH8PbblYsrhDyKqtNGOPpon8vm+9/3KlTv3vCHP8Cbb8KJJ8LEid7bOITNXSSajdCvH5x/vrfV3H67j/oeORK22w7+679gyRKvYnXpUulIQ6isSDQbaZdd/PHWW1BfDyNGeMK55Ra48054+WU44wyfCD2EzVUkmk6y9daNyxKccIIPzrzmGjj9dPjUp/yumGPGNCadlSvhj3+EffeNRBQ+2lqcYS/POnuGvSwtXw433wzPPAPLlkFVlfe92XNPmDoV6uo8MR14IAwYAAsX+kx/Rx65YZVr8WIYOBC6xs9DqJCOzrAXiWYTWroUfvGLxuknamvhtNNg5ky47z5Yv95fe+stn/XvzDN9DuO6OrjpJnj0UR9Vfu653q8nhE0tEk2B/PWv8PTTcMwxUFPjr733npdUqqv99i/XXOOln+pqWL3aSzf77+9JavBg+NKXYNEi78szciTsuKOPyaqu9jtwLlsG/fv7egidJRLNR8xLL8HDD3vDcvfucOih3g707LM+X04p+XTp4smmpLq6cb1PH99v6FBPbnV1fsl93329uvb++57Mqqo80T33HPzjH57Ihg/30lXJqlUeRySuzVskms3I8uV+H6rBg319/nwf6Ll8uSeEPn380vtTT/mUpGY+KLRXL+/js8MOnmQWLvS/9enjV8zWr//w+2y3nbclzZsHzz/vx/jGN7xktWiRv2+vXv5eQ4ZAt27pP8O6dZ7gSu1Na9Z4MtxnH+jRo/V9wTtGzp0Ls2Z558kJE/x47bVsmccf/Z3SiUQTmrVkiSegHXf0pDJtmrcH9e8P22/vJaIVK7yz4ZgxXvpZvBheecVP/Bde8EGln/yk35bm5Zd9YrDVTW7zV10No0Z59a2hwRNJQ4Mnko9/3B/vvAMLFsCLL3qS6NULjjrKk9TVV3vb1KBB8J3veMK8915Pfnvv7V0Ievf2950+HR580I9XMny4TzD/7LOeTPfeG8aN878tXepx1NR494MBAzxR/fzncM89sNtuPjF9jx7+eUtDTbbaqv3f9Ysv+mfddtuO/5uZee/y/v03TIBm/iglVTP/Trt188+1bp1/Xsm/y84ugUaiCZl4/30/8SQ/Of/7v71aN2qUJ6/Vq/2kmDPHT9L6ev/P3a2bP95778NTalRV+WX+0aP99jbPPeevb7utV+t+8xs/UcBPtF69NpySo6rKk8r48V7ieuYZ7yBZV+ftVQMHwpNPbpgMwT/HXnt5EnzmGU+gzz/via2pIUNg2DCPraHBj1dd7Ym2Vy9vkF+71k/0l1/2El7Jrrt60m5o8Crnttt6yWnNGv+OVqzw7g2lUmSPHl5VffddmDHDk+6wYX41Evw9Fi7076KqCvbbz7/HP/3JS5zNkfyY3bv78XfYwRN2ba3/20iNPwgNDf7v26+ff+/r1nmMb77p711f768feWQkmpBTK1Z4Cal/f6/ulVexZs/2k+fgg/1kqK/3YRwDB3qJpGtXH0v26quetMy8tNK0tLF+vZ8cPZM7w69bB3/7m68PGOB/X77cq5N//KMnlpNP9vFq9fX+Wq9eniDWrfMq2Zw5fnIvWeIna8+efkLW13/43l+lktLYsbDzzp7AHnzQE0m3bp6g1qz5cLyS79etmx+rvt63kbw0NXq0V3tfesm379fPE9fQof49PPaYbz90KBx+uMe3dKknwlKJrRT7unWewObO3TCONKqq/LPX18O99xY80UjqBZwPLALeNLM7W9o2Ek3YGA0NngT690+3vdmHqzBmXpJ57z0v7fXt2/b+pXa1Ummod+8Nq0WlBNa7d+Nrb73l+zS9VfPatV7aGDYsffvS+vXelrdypZdUzRpLnqUSzrJl/uje3ac/qa1t/HFYswZ69ix+ojkGWGdmd0j6nZlNbPL3ScAkgKFDh+65YMGCSoQZwmat02+JWwFDgbpkeYO7X5vZDWY21szG1pZfdw0h5F6eEs3rQCmDNNOMF0IoqjyNmrkLmCxpIHBrpYMJIXSe3CQaM3sPOLPScYQQOl+eqk4hhI+oSDQhhMxFogkhZC43/WjaQ9IqoGh3wx4ALK10EO1QtHghYt4URppZn7Y3+7DcNAa309yOdBqqJEkzixRz0eKFiHlTkNShLvlRdQohZC4STQghc0VNNDdUOoAOKFrMRYsXIuZNoUPxFrIxOIRQLEUt0YQQCiQSTQghc5Fowv+S1EXSXEnbVTqWUHmSuko6R9JGtyMVqh9Ne2bhqzRJnwd2AboAyYSMDAQGA1PMrAOTKmbuKPy7RdLJQAMwzMzOq2hULZDUBTgdn8doK+ANcv4dS9oD+L/AE3jM88lvzL2AacBJAJK+QlmseEEl1flYtBLNRGCWmV0NHFvpYNrwjJldAlyNn8DHJnE/jX+OXJHUDRiEzwtUBXzazG4EVkrat6LBtexwfMK0auApcv4dJ+YBPYGP4zHmNmYzW8mHey03jTX1+Vi0RNPqLHx5YmZvJItfBC4DSncrqgOGVSSo1h0D/L9k2YD6ZDmv8YKfrP8ws58D55L/7xg8Od4HnEVxYi5pGmvq87FoiaZQs/BJ+gz+C7YIKBWJa4GFFQuqZbXAQcAOwFeBvmWv5zFegLeA0o1Sqsj/dww+tmmZmX0AnEMxYi5pGmvq87FQ/WiSNprJ+IdcYmZ3VDailkk6FK+/PouftHfj/yhDgMlmlrtEKemfgP/ESzYfAGuBEWb2/YoG1gJJNcAlwGygG17Mz/t3PAg4G3gZGAH8lRzHLOks4DDgVGAkZbECIuX5WKhEE0IopqJVnUIIBRSJJoSQuUg0IYTMRaIJIWQuEk0IIXORaDqZpPGSZkh6POlti6TtJN0s6W5Ju3bCe0yVNGOjg033Xp+WdKOkyyQdl2L70ZKmSzphE4QXCiISTSczsz8DM4D1wJXJa4uAm4G7zey5TnibazrhGGn9C3CbmX2Xxp7DLTKzF4A/Zx5VKJRCDaosmP8D3CbpeDP7RfkfJF0CfMLMJkg6ErjBzLZMxhRdCzyC/9vsDvwHcAAwFjjVzEqTQ9dI+i6wLT7Q7RtmtjYZDDkS77zWF7/752HAFcAf8EGeE81sSJOYJgE7AcvxTlnfAfYD9gK6SdrGzG4r2/544KfAj4A+wG7A6WY2r8lxTwZ2Bpbg3dZPAiYAtwE/M7Mpks4HPgkcjSe2NuPHezH/O/AiMBq40MxeKnvfI5L4bsd7tI4FLjCzpyXtjA8BeB4fxnAR8E6yrfAOgJ9Ntr+77Jhj8MGEM/Hu+P8CnAa8AFwOPI4PpP2xmc2W9D3gAuDfknhX4z9CB+ADKo8ws/WSxgFfTz7LSOBsMyvSnRHaZmbx6OQH3ltyON7z83U8YUwATkj+PhyYUbb9/Cb7Xpgs/yvwy2R5InBV2f4Lgapk/Vp8RPAoYA6NHTFvBr5QtvztZHlsk3hHAc+VrV8LTCrbb0ILn3M+sEOyfBRwR9lnKH3Ww8vi/Cnw+WT5B8CZyfLZyWdKHX/y3VyPD6gcAWzbTHwzgIOS5X2Ap5LlJ4BPJssTgN+VLT+ZLG8DDG5yvKeAfZLlg0r/hsB2wJhkeUzpeyj7jj6WLM8GDkyWfw/siSe2xaX3Ak4Arqj0/+HOfkSJJkNmNi9pq/gN/gua1qvJ8/Ky5XfwkkPJfPPxMgCv4KWGt/ChA2dJAp/moaZsnzlJXE1vmTEaPyFKXsFLKGm81iSGpt4DpkpaCvwTPvIXPJk9LOmn+Ek2P5mGIFX8kubgCeoR/B5f32lHfLsCh0gaj4+kfreZ9/hHM8faGR86UH7cUpxHJ8NOamgc/0NyrNb+PQfgpZvjks+8FV7t/kiJRJMxM3tQ0nX4r3lpXpdVJCeQpB40+Y+Z0jBJVUmy2QmvBjwPrDafnqJU1G8oD6eFYz2PlwpKdgRmpYxje/wk3gkv+jf1W2A3M1uYjE3yQMzelPQk8EvgqrI40sa/D3CJmZ0n6T+A4/HqS3PxvdYkvmeBu8zsOUnd+fD0DK2NyXkxOc5fkuOWnA28Y2YXSdopiS2tpfgI6OvN7B1J/YG8TsvRYZFoOpmkTwHjgd6SLjCzejO7XNLupW3M7G1JsyX9CD8JVkg6CZie7LuLpMfxake/5D/vccCukvYGvoxP43C+pK3wX8YbzWyNpOslXY4ns/7A2ck+u+K/movM7JXymM3s75KuknQlsAJYB9yUfJbSfivN7Gk2dJCkE4E9gFMljS77DNOA64CfSXoUP4FGSrrPzOqAn+DtU38uiyNt/FsBl0t6DU/ULTWQ75W0gexNMoET8A3g3yTNwwcI/ipJOKXv+J+t+UmcTgIuTBLkWhqT0p3AxckxqvEfgQOTGPsmpdoFeBvVCZLuKX0evBp3NHCppNfxqReaS5iFFoMqQ4dJmm9mwzuwXxV+xXNPYJSZ3dzJoZXeZwbeVjS/k463BzAnSejjgePN7JudceyPuri8HTpE0jH4r/W3O7D7DsAd+C/6bW1s2yHJVKrDgI7E15LRwNXJ1aTj8CtQIYUo0YQQMhclmhBC5iLRhBAyF4kmhJC5SDQhhMxFogkhZO5/AKp6sHqIDebDAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize = (4, 3))\n",
    "ax.plot(x, y, color = \"blue\", alpha = 0.7)\n",
    "ax.set_xlabel(\"Number of players per game\")\n",
    "ax.set_ylabel(\"Number of games\")\n",
    "ax.set_xlim([0,100])\n",
    "fig.savefig(\"c1.svg\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We choose `c1` = 10, based on intuition."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "filter_appid = groupby_appid[groupby_appid > 10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "filter_appid_df = pd.merge(filter_appid.reset_index(), df, on = \"appid\").drop(\"steamid_x\", axis = 1).rename(index = int, columns = {\"steamid_y\" : \"steamid\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>appid</th>\n",
       "      <th>steamid</th>\n",
       "      <th>playtime_forever</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10</td>\n",
       "      <td>76561197969918174</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10</td>\n",
       "      <td>76561198001033415</td>\n",
       "      <td>582</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10</td>\n",
       "      <td>76561197960794555</td>\n",
       "      <td>251</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10</td>\n",
       "      <td>76561198053398526</td>\n",
       "      <td>211</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10</td>\n",
       "      <td>76561197967925129</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   appid            steamid  playtime_forever\n",
       "0     10  76561197969918174                 3\n",
       "1     10  76561198001033415               582\n",
       "2     10  76561197960794555               251\n",
       "3     10  76561198053398526               211\n",
       "4     10  76561197967925129                19"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filter_appid_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unique SteamIds : 15582\n",
      "Unique AppIds : 12982\n",
      "Density : 1.65 %\n"
     ]
    }
   ],
   "source": [
    "print(\"Unique SteamIds :\", filter_appid_df[\"steamid\"].nunique())\n",
    "print(\"Unique AppIds :\", filter_appid_df[\"appid\"].nunique())\n",
    "print(\"Density : {0:.2f} %\".format(density(filter_appid_df) * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By removing sparse games, the density of the new data has improved. The unique `appids` have decreased naturally while the unique `steamids` have not decreased. We now perform a similar procedure that discards the sparse players."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "groupby_steamid = filter_appid_df.groupby(by = \"steamid\").count()[\"appid\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### We plot the distribtion of `appid` counts to determine the cutoff `c2`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "groupby_steamid_nparr = groupby_steamid.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize = (4, 3))#ncols = 2, figsize = (14, 4))\n",
    "counts, bin_edges = np.histogram(groupby_steamid_nparr, bins = 2000, density = True)\n",
    "ecdf = np.cumsum(counts)\n",
    "#ax.hist(groupby_steamid_nparr, density = True, bins = 2000, color = \"blue\", alpha = 0.7)\n",
    "#ax.set_xlabel(\"Unique AppIDs\")\n",
    "#ax.set_xlim([0, 500])\n",
    "#ax.set_ylabel(\"PDF\")\n",
    "ax.plot(bin_edges[1:], ecdf/ecdf[-1], color = \"blue\", alpha = 0.7)\n",
    "ax.set_xlim([0, 500])\n",
    "ax.set_xlabel(\"Unique AppIDs\")\n",
    "ax.set_ylabel(\"CDF\")\n",
    "#ax.set_title(\"CDF\")\n",
    "fig.savefig(\"c2_cdf.svg\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We determine `c2` to be 65, the reason being that the distribution of `appid` counts start to increase rapidly after 65 (users with number of unique games less than 65 are rare). We want to discard such users because they are quite sparse compared to the rest of the data and result in negligible data loss but noticably increase the density."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "filter_steamid = groupby_steamid[groupby_steamid > 35]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "filter_steamid_df = pd.merge(filter_steamid.reset_index(), filter_appid_df, on = \"steamid\").drop(\"appid_x\", axis = 1).rename(index = int, columns = {\"appid_y\" : \"appid\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unique SteamIds : 13845\n",
      "Unique AppIds : 12982\n",
      "Density : 1.84 %\n"
     ]
    }
   ],
   "source": [
    "print(\"Unique SteamIds :\", filter_steamid_df[\"steamid\"].nunique())\n",
    "print(\"Unique AppIds :\", filter_steamid_df[\"appid\"].nunique())\n",
    "print(\"Density : {0:.2f} %\".format(density(filter_steamid_df) * 100))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dumping the sampled data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "#filter_steamid_df.to_csv(\"final_data.csv\")"
   ]
  }
 ],
 "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}