AS_8_cumulative_categorical_features.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>genre_id</th>\n",
       "      <th>ts_listen</th>\n",
       "      <th>album_id</th>\n",
       "      <th>context_type</th>\n",
       "      <th>platform_name</th>\n",
       "      <th>listen_type</th>\n",
       "      <th>user_id</th>\n",
       "      <th>artist_id</th>\n",
       "      <th>is_listened</th>\n",
       "      <th>ts_listen_dayofweek</th>\n",
       "      <th>ts_listen_hour</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7578747</th>\n",
       "      <td>0</td>\n",
       "      <td>1479664431</td>\n",
       "      <td>14226364</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>12641</td>\n",
       "      <td>9285528</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578748</th>\n",
       "      <td>0</td>\n",
       "      <td>1480358099</td>\n",
       "      <td>14226364</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>10055</td>\n",
       "      <td>9285528</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578749</th>\n",
       "      <td>0</td>\n",
       "      <td>1480616853</td>\n",
       "      <td>14581358</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1029</td>\n",
       "      <td>129</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578750</th>\n",
       "      <td>0</td>\n",
       "      <td>1479644510</td>\n",
       "      <td>14581358</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4630</td>\n",
       "      <td>129</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578751</th>\n",
       "      <td>0</td>\n",
       "      <td>1480491885</td>\n",
       "      <td>14581358</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>6467</td>\n",
       "      <td>129</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         genre_id   ts_listen  album_id  context_type  platform_name  \\\n",
       "7578747         0  1479664431  14226364             5              0   \n",
       "7578748         0  1480358099  14226364             1              0   \n",
       "7578749         0  1480616853  14581358             1              0   \n",
       "7578750         0  1479644510  14581358             1              0   \n",
       "7578751         0  1480491885  14581358             1              1   \n",
       "\n",
       "         listen_type  user_id  artist_id  is_listened  ts_listen_dayofweek  \\\n",
       "7578747            1    12641    9285528            0                    6   \n",
       "7578748            1    10055    9285528            0                    0   \n",
       "7578749            1     1029        129            0                    3   \n",
       "7578750            1     4630        129            0                    6   \n",
       "7578751            1     6467        129            0                    2   \n",
       "\n",
       "         ts_listen_hour  \n",
       "7578747              17  \n",
       "7578748              18  \n",
       "7578749              18  \n",
       "7578750              12  \n",
       "7578751               7  "
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "path_to_train = './data/train.csv'\n",
    "path_to_test = './data/test.csv'\n",
    "\n",
    "categorical_features = ['listen_type', \n",
    "                        'genre_id', \n",
    "                        'album_id', \n",
    "                        'context_type',\n",
    "                        'platform_name',\n",
    "                        'artist_id']\n",
    "\n",
    "train = pd.read_csv(path_to_train, encoding='utf-8', usecols=['user_id', 'ts_listen', 'is_listened'] + categorical_features)\n",
    "test = pd.read_csv(path_to_test, encoding='utf-8', usecols=['user_id', 'ts_listen'] + categorical_features)\n",
    "test['is_listened'] = 0 # просто щоб зібрати, ми будемо використовувати попередні значення\n",
    "\n",
    "path_to_test_ts_listen = './data/features/ts_listen.test.csv'\n",
    "path_to_train_ts_listen = './data/features/ts_listen.train.csv'\n",
    "categorical_features_time = ['ts_listen_dayofweek', \n",
    "                            'ts_listen_hour']\n",
    "\n",
    "test_ts_listen = pd.read_csv(path_to_test_ts_listen, encoding='utf-8', usecols=categorical_features_time)\n",
    "train_ts_listen = pd.read_csv(path_to_train_ts_listen, encoding='utf-8', usecols=categorical_features_time)\n",
    "\n",
    "test = pd.concat([test, test_ts_listen], axis=1)\n",
    "train = pd.concat([train, train_ts_listen], axis=1)\n",
    "\n",
    "df = pd.concat([train, test], ignore_index=True)\n",
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def mean_cum_by_category_by_user(df, category):\n",
    "    df.sort_values(by=['user_id', 'ts_listen', category], axis=0, inplace=True)\n",
    "    \n",
    "    is_listened_cumsum = df.groupby(['user_id',category]).agg({'is_listened':'cumsum'})['is_listened']\n",
    "    \n",
    "    cumcount = df.groupby(['user_id',category]).cumcount() + 1\n",
    "    \n",
    "    column_name = category+\"_cumulative_mean\"\n",
    "    \n",
    "    df[column_name] = (is_listened_cumsum/cumcount).shift()\n",
    "    \n",
    "    fi = df.groupby(['user_id',category], as_index=False).apply(lambda x: x.index[0])\n",
    "    \n",
    "    df.ix[fi, column_name] = 0\n",
    "    \n",
    "    return column_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>genre_id</th>\n",
       "      <th>ts_listen</th>\n",
       "      <th>album_id</th>\n",
       "      <th>context_type</th>\n",
       "      <th>platform_name</th>\n",
       "      <th>listen_type</th>\n",
       "      <th>user_id</th>\n",
       "      <th>artist_id</th>\n",
       "      <th>is_listened</th>\n",
       "      <th>ts_listen_dayofweek</th>\n",
       "      <th>ts_listen_hour</th>\n",
       "      <th>listen_type_cumulative_mean</th>\n",
       "      <th>genre_id_cumulative_mean</th>\n",
       "      <th>album_id_cumulative_mean</th>\n",
       "      <th>context_type_cumulative_mean</th>\n",
       "      <th>platform_name_cumulative_mean</th>\n",
       "      <th>artist_id_cumulative_mean</th>\n",
       "      <th>ts_listen_dayofweek_cumulative_mean</th>\n",
       "      <th>ts_listen_hour_cumulative_mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>682493</th>\n",
       "      <td>0</td>\n",
       "      <td>1477939775</td>\n",
       "      <td>12811130</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7508018</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4454939</th>\n",
       "      <td>0</td>\n",
       "      <td>1477982531</td>\n",
       "      <td>12920172</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1035779</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5354523</th>\n",
       "      <td>0</td>\n",
       "      <td>1477982561</td>\n",
       "      <td>13947748</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4331004</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5899458</th>\n",
       "      <td>0</td>\n",
       "      <td>1477982582</td>\n",
       "      <td>13976722</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3469</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4716574</th>\n",
       "      <td>0</td>\n",
       "      <td>1477982792</td>\n",
       "      <td>13640610</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>282118</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         genre_id   ts_listen  album_id  context_type  platform_name  \\\n",
       "682493          0  1477939775  12811130             0              0   \n",
       "4454939         0  1477982531  12920172             0              0   \n",
       "5354523         0  1477982561  13947748             0              0   \n",
       "5899458         0  1477982582  13976722             0              0   \n",
       "4716574         0  1477982792  13640610             0              0   \n",
       "\n",
       "         listen_type  user_id  artist_id  is_listened  ts_listen_dayofweek  \\\n",
       "682493             0        0    7508018            1                    0   \n",
       "4454939            0        0    1035779            0                    1   \n",
       "5354523            0        0    4331004            0                    1   \n",
       "5899458            0        0       3469            1                    1   \n",
       "4716574            0        0     282118            1                    1   \n",
       "\n",
       "         ts_listen_hour  listen_type_cumulative_mean  \\\n",
       "682493               18                     0.000000   \n",
       "4454939               6                     1.000000   \n",
       "5354523               6                     0.500000   \n",
       "5899458               6                     0.333333   \n",
       "4716574               6                     0.500000   \n",
       "\n",
       "         genre_id_cumulative_mean  album_id_cumulative_mean  \\\n",
       "682493                   0.000000                       0.0   \n",
       "4454939                  1.000000                       0.0   \n",
       "5354523                  0.500000                       0.0   \n",
       "5899458                  0.333333                       0.0   \n",
       "4716574                  0.500000                       0.0   \n",
       "\n",
       "         context_type_cumulative_mean  platform_name_cumulative_mean  \\\n",
       "682493                       0.000000                       0.000000   \n",
       "4454939                      1.000000                       1.000000   \n",
       "5354523                      0.500000                       0.500000   \n",
       "5899458                      0.333333                       0.333333   \n",
       "4716574                      0.500000                       0.500000   \n",
       "\n",
       "         artist_id_cumulative_mean  ts_listen_dayofweek_cumulative_mean  \\\n",
       "682493                         0.0                             0.000000   \n",
       "4454939                        0.0                             0.000000   \n",
       "5354523                        0.0                             0.000000   \n",
       "5899458                        0.0                             0.000000   \n",
       "4716574                        0.0                             0.333333   \n",
       "\n",
       "         ts_listen_hour_cumulative_mean  \n",
       "682493                         0.000000  \n",
       "4454939                        0.000000  \n",
       "5354523                        0.000000  \n",
       "5899458                        0.000000  \n",
       "4716574                        0.333333  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_cols = [mean_cum_by_category_by_user(df, category) for category in categorical_features+categorical_features_time]\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>genre_id</th>\n",
       "      <th>ts_listen</th>\n",
       "      <th>album_id</th>\n",
       "      <th>context_type</th>\n",
       "      <th>platform_name</th>\n",
       "      <th>listen_type</th>\n",
       "      <th>user_id</th>\n",
       "      <th>artist_id</th>\n",
       "      <th>is_listened</th>\n",
       "      <th>ts_listen_dayofweek</th>\n",
       "      <th>ts_listen_hour</th>\n",
       "      <th>listen_type_cumulative_mean</th>\n",
       "      <th>genre_id_cumulative_mean</th>\n",
       "      <th>album_id_cumulative_mean</th>\n",
       "      <th>context_type_cumulative_mean</th>\n",
       "      <th>platform_name_cumulative_mean</th>\n",
       "      <th>artist_id_cumulative_mean</th>\n",
       "      <th>ts_listen_dayofweek_cumulative_mean</th>\n",
       "      <th>ts_listen_hour_cumulative_mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7578747</th>\n",
       "      <td>0</td>\n",
       "      <td>1479664431</td>\n",
       "      <td>14226364</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>12641</td>\n",
       "      <td>9285528</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>17</td>\n",
       "      <td>0.513514</td>\n",
       "      <td>0.176471</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.208696</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.428571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578748</th>\n",
       "      <td>0</td>\n",
       "      <td>1480358099</td>\n",
       "      <td>14226364</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>10055</td>\n",
       "      <td>9285528</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>18</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.967742</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.970588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578749</th>\n",
       "      <td>0</td>\n",
       "      <td>1480616853</td>\n",
       "      <td>14581358</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1029</td>\n",
       "      <td>129</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>18</td>\n",
       "      <td>0.822097</td>\n",
       "      <td>0.911032</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.822097</td>\n",
       "      <td>0.875566</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.843902</td>\n",
       "      <td>0.950000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578750</th>\n",
       "      <td>0</td>\n",
       "      <td>1479644510</td>\n",
       "      <td>14581358</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4630</td>\n",
       "      <td>129</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "      <td>0.904372</td>\n",
       "      <td>0.900000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.840580</td>\n",
       "      <td>0.846914</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.880952</td>\n",
       "      <td>0.878049</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578751</th>\n",
       "      <td>0</td>\n",
       "      <td>1480491885</td>\n",
       "      <td>14581358</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>6467</td>\n",
       "      <td>129</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>0.932203</td>\n",
       "      <td>0.887324</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.932203</td>\n",
       "      <td>0.925532</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.928571</td>\n",
       "      <td>0.925926</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         genre_id   ts_listen  album_id  context_type  platform_name  \\\n",
       "7578747         0  1479664431  14226364             5              0   \n",
       "7578748         0  1480358099  14226364             1              0   \n",
       "7578749         0  1480616853  14581358             1              0   \n",
       "7578750         0  1479644510  14581358             1              0   \n",
       "7578751         0  1480491885  14581358             1              1   \n",
       "\n",
       "         listen_type  user_id  artist_id  is_listened  ts_listen_dayofweek  \\\n",
       "7578747            1    12641    9285528            0                    6   \n",
       "7578748            1    10055    9285528            0                    0   \n",
       "7578749            1     1029        129            0                    3   \n",
       "7578750            1     4630        129            0                    6   \n",
       "7578751            1     6467        129            0                    2   \n",
       "\n",
       "         ts_listen_hour  listen_type_cumulative_mean  \\\n",
       "7578747              17                     0.513514   \n",
       "7578748              18                     0.750000   \n",
       "7578749              18                     0.822097   \n",
       "7578750              12                     0.904372   \n",
       "7578751               7                     0.932203   \n",
       "\n",
       "         genre_id_cumulative_mean  album_id_cumulative_mean  \\\n",
       "7578747                  0.176471                       0.0   \n",
       "7578748                  1.000000                       1.0   \n",
       "7578749                  0.911032                       0.0   \n",
       "7578750                  0.900000                       0.0   \n",
       "7578751                  0.887324                       0.0   \n",
       "\n",
       "         context_type_cumulative_mean  platform_name_cumulative_mean  \\\n",
       "7578747                      0.000000                       0.208696   \n",
       "7578748                      1.000000                       0.967742   \n",
       "7578749                      0.822097                       0.875566   \n",
       "7578750                      0.840580                       0.846914   \n",
       "7578751                      0.932203                       0.925532   \n",
       "\n",
       "         artist_id_cumulative_mean  ts_listen_dayofweek_cumulative_mean  \\\n",
       "7578747                        0.0                             0.200000   \n",
       "7578748                        1.0                             1.000000   \n",
       "7578749                        0.0                             0.843902   \n",
       "7578750                        0.0                             0.880952   \n",
       "7578751                        0.0                             0.928571   \n",
       "\n",
       "         ts_listen_hour_cumulative_mean  \n",
       "7578747                        0.428571  \n",
       "7578748                        0.970588  \n",
       "7578749                        0.950000  \n",
       "7578750                        0.878049  \n",
       "7578751                        0.925926  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sort_index(inplace=True)\n",
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "path_to_test_cum_mean = './data/features/cum_mean.test.csv'\n",
    "path_to_train_cum_mean = './data/features/cum_mean.train.csv'\n",
    "\n",
    "test_cum_mean = df.iloc[range(len(train), len(train)+len(test))][new_cols]\n",
    "test_cum_mean.to_csv(path_to_test_cum_mean, encoding='utf-8', index=False)\n",
    "\n",
    "train_cum_mean = df.iloc[range(len(train))][new_cols]\n",
    "train_cum_mean.to_csv(path_to_train_cum_mean, encoding='utf-8', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "------\n",
    "\n",
    "More features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ts_listen</th>\n",
       "      <th>media_id</th>\n",
       "      <th>user_id</th>\n",
       "      <th>is_listened</th>\n",
       "      <th>release_date_year</th>\n",
       "      <th>release_date_decade</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7578747</th>\n",
       "      <td>1479664431</td>\n",
       "      <td>133549164</td>\n",
       "      <td>12641</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578748</th>\n",
       "      <td>1480358099</td>\n",
       "      <td>133549164</td>\n",
       "      <td>10055</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578749</th>\n",
       "      <td>1480616853</td>\n",
       "      <td>136334560</td>\n",
       "      <td>1029</td>\n",
       "      <td>0</td>\n",
       "      <td>1991</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578750</th>\n",
       "      <td>1479644510</td>\n",
       "      <td>136334560</td>\n",
       "      <td>4630</td>\n",
       "      <td>0</td>\n",
       "      <td>1991</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578751</th>\n",
       "      <td>1480491885</td>\n",
       "      <td>136334560</td>\n",
       "      <td>6467</td>\n",
       "      <td>0</td>\n",
       "      <td>1991</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          ts_listen   media_id  user_id  is_listened  release_date_year  \\\n",
       "7578747  1479664431  133549164    12641            0               2016   \n",
       "7578748  1480358099  133549164    10055            0               2016   \n",
       "7578749  1480616853  136334560     1029            0               1991   \n",
       "7578750  1479644510  136334560     4630            0               1991   \n",
       "7578751  1480491885  136334560     6467            0               1991   \n",
       "\n",
       "         release_date_decade  \n",
       "7578747                   11  \n",
       "7578748                   11  \n",
       "7578749                    9  \n",
       "7578750                    9  \n",
       "7578751                    9  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "path_to_train = './data/train.csv'\n",
    "path_to_test = './data/test.csv'\n",
    "\n",
    "train = pd.read_csv(path_to_train, encoding='utf-8', usecols=['user_id', 'ts_listen', 'media_id', 'is_listened'])\n",
    "test = pd.read_csv(path_to_test, encoding='utf-8', usecols=['user_id', 'ts_listen', 'media_id'])\n",
    "test['is_listened'] = 0 # просто щоб зібрати, ми будемо використовувати попередні значення\n",
    "\n",
    "path_to_test_release_date = './data/features/release_date.test.txt'\n",
    "path_to_train_release_date = './data/features/release_date.train.txt'\n",
    "categorical_features_release_date = ['release_date_decade', 'release_date_year']\n",
    "\n",
    "test_release_date = pd.read_csv(path_to_test_release_date, encoding='utf-8', usecols=categorical_features_release_date)\n",
    "train_release_date = pd.read_csv(path_to_train_release_date, encoding='utf-8', usecols=categorical_features_release_date)\n",
    "\n",
    "test = pd.concat([test, test_release_date], axis=1)\n",
    "train = pd.concat([train, train_release_date], axis=1)\n",
    "\n",
    "df = pd.concat([train, test], ignore_index=True)\n",
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/tolik/.local/lib/python3.5/site-packages/IPython/core/interactiveshell.py:2683: DtypeWarning: Columns (10) have mixed types. Specify dtype option on import or set low_memory=False.\n",
      "  interactivity=interactivity, compiler=compiler, result=result)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ts_listen</th>\n",
       "      <th>user_id</th>\n",
       "      <th>is_listened</th>\n",
       "      <th>release_date_year</th>\n",
       "      <th>release_date_decade</th>\n",
       "      <th>ExplicitLyrics</th>\n",
       "      <th>Isrc_Country</th>\n",
       "      <th>Isrc_Label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1480597215</td>\n",
       "      <td>9241</td>\n",
       "      <td>0</td>\n",
       "      <td>2004</td>\n",
       "      <td>10</td>\n",
       "      <td>False</td>\n",
       "      <td>FR</td>\n",
       "      <td>6V8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1480544735</td>\n",
       "      <td>16547</td>\n",
       "      <td>1</td>\n",
       "      <td>2006</td>\n",
       "      <td>10</td>\n",
       "      <td>False</td>\n",
       "      <td>FR</td>\n",
       "      <td>6V8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1479563953</td>\n",
       "      <td>7665</td>\n",
       "      <td>1</td>\n",
       "      <td>2014</td>\n",
       "      <td>11</td>\n",
       "      <td>False</td>\n",
       "      <td>FR</td>\n",
       "      <td>6V8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1480152098</td>\n",
       "      <td>1580</td>\n",
       "      <td>0</td>\n",
       "      <td>2000</td>\n",
       "      <td>10</td>\n",
       "      <td>False</td>\n",
       "      <td>NL</td>\n",
       "      <td>A20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1478368974</td>\n",
       "      <td>1812</td>\n",
       "      <td>1</td>\n",
       "      <td>2008</td>\n",
       "      <td>10</td>\n",
       "      <td>False</td>\n",
       "      <td>US</td>\n",
       "      <td>SM1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    ts_listen  user_id  is_listened  release_date_year  release_date_decade  \\\n",
       "0  1480597215     9241            0               2004                   10   \n",
       "1  1480544735    16547            1               2006                   10   \n",
       "2  1479563953     7665            1               2014                   11   \n",
       "3  1480152098     1580            0               2000                   10   \n",
       "4  1478368974     1812            1               2008                   10   \n",
       "\n",
       "  ExplicitLyrics Isrc_Country Isrc_Label  \n",
       "0          False           FR        6V8  \n",
       "1          False           FR        6V8  \n",
       "2          False           FR        6V8  \n",
       "3          False           NL        A20  \n",
       "4          False           US        SM1  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# track info categorical features\n",
    "path_to_test_release_date = './data/features/all_media_info.csv'\n",
    "track_info = pd.read_csv(path_to_test_release_date, encoding='utf-8', usecols=['Id', 'ExplicitLyrics', 'Isrc_Country', 'Isrc_Label'])\n",
    "\n",
    "df = pd.merge(df, track_info, left_on='media_id', right_on='Id', how='left', sort=False)\n",
    "del df['Id']\n",
    "del df['media_id']\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/tolik/.local/lib/python3.5/site-packages/ipykernel_launcher.py:14: DeprecationWarning: \n",
      ".ix is deprecated. Please use\n",
      ".loc for label based indexing or\n",
      ".iloc for positional indexing\n",
      "\n",
      "See the documentation here:\n",
      "http://pandas.pydata.org/pandas-docs/stable/indexing.html#deprecate_ix\n",
      "  \n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ts_listen</th>\n",
       "      <th>user_id</th>\n",
       "      <th>is_listened</th>\n",
       "      <th>release_date_year</th>\n",
       "      <th>release_date_decade</th>\n",
       "      <th>ExplicitLyrics</th>\n",
       "      <th>Isrc_Country</th>\n",
       "      <th>Isrc_Label</th>\n",
       "      <th>release_date_year_cumulative_mean</th>\n",
       "      <th>release_date_decade_cumulative_mean</th>\n",
       "      <th>ExplicitLyrics_cumulative_mean</th>\n",
       "      <th>Isrc_Country_cumulative_mean</th>\n",
       "      <th>Isrc_Label_cumulative_mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>682493</th>\n",
       "      <td>1477939775</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>False</td>\n",
       "      <td>DK</td>\n",
       "      <td>4YA</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4454939</th>\n",
       "      <td>1477982531</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>False</td>\n",
       "      <td>FR</td>\n",
       "      <td>UM7</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5354523</th>\n",
       "      <td>1477982561</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>False</td>\n",
       "      <td>US</td>\n",
       "      <td>SM1</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5899458</th>\n",
       "      <td>1477982582</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>False</td>\n",
       "      <td>US</td>\n",
       "      <td>RC1</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4716574</th>\n",
       "      <td>1477982792</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>False</td>\n",
       "      <td>QM</td>\n",
       "      <td>UY4</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          ts_listen  user_id  is_listened  release_date_year  \\\n",
       "682493   1477939775        0            1               2016   \n",
       "4454939  1477982531        0            0               2016   \n",
       "5354523  1477982561        0            0               2016   \n",
       "5899458  1477982582        0            1               2016   \n",
       "4716574  1477982792        0            1               2016   \n",
       "\n",
       "         release_date_decade ExplicitLyrics Isrc_Country Isrc_Label  \\\n",
       "682493                    11          False           DK        4YA   \n",
       "4454939                   11          False           FR        UM7   \n",
       "5354523                   11          False           US        SM1   \n",
       "5899458                   11          False           US        RC1   \n",
       "4716574                   11          False           QM        UY4   \n",
       "\n",
       "         release_date_year_cumulative_mean  \\\n",
       "682493                            0.000000   \n",
       "4454939                           1.000000   \n",
       "5354523                           0.500000   \n",
       "5899458                           0.333333   \n",
       "4716574                           0.500000   \n",
       "\n",
       "         release_date_decade_cumulative_mean  ExplicitLyrics_cumulative_mean  \\\n",
       "682493                              0.000000                        0.000000   \n",
       "4454939                             1.000000                        1.000000   \n",
       "5354523                             0.500000                        0.500000   \n",
       "5899458                             0.333333                        0.333333   \n",
       "4716574                             0.500000                        0.500000   \n",
       "\n",
       "         Isrc_Country_cumulative_mean  Isrc_Label_cumulative_mean  \n",
       "682493                            0.0                         0.0  \n",
       "4454939                           0.0                         0.0  \n",
       "5354523                           0.0                         0.0  \n",
       "5899458                           0.0                         0.0  \n",
       "4716574                           0.0                         0.0  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols = ['release_date_year', 'release_date_decade', 'ExplicitLyrics', 'Isrc_Country', 'Isrc_Label']\n",
    "\n",
    "new_cols = [mean_cum_by_category_by_user(df, category) for category in cols]\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ts_listen</th>\n",
       "      <th>user_id</th>\n",
       "      <th>is_listened</th>\n",
       "      <th>release_date_year</th>\n",
       "      <th>release_date_decade</th>\n",
       "      <th>ExplicitLyrics</th>\n",
       "      <th>Isrc_Country</th>\n",
       "      <th>Isrc_Label</th>\n",
       "      <th>release_date_year_cumulative_mean</th>\n",
       "      <th>release_date_decade_cumulative_mean</th>\n",
       "      <th>ExplicitLyrics_cumulative_mean</th>\n",
       "      <th>Isrc_Country_cumulative_mean</th>\n",
       "      <th>Isrc_Label_cumulative_mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7578747</th>\n",
       "      <td>1479664431</td>\n",
       "      <td>12641</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>False</td>\n",
       "      <td>GB</td>\n",
       "      <td>2LD</td>\n",
       "      <td>0.116667</td>\n",
       "      <td>0.190000</td>\n",
       "      <td>0.224490</td>\n",
       "      <td>0.288462</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578748</th>\n",
       "      <td>1480358099</td>\n",
       "      <td>10055</td>\n",
       "      <td>0</td>\n",
       "      <td>2016</td>\n",
       "      <td>11</td>\n",
       "      <td>False</td>\n",
       "      <td>GB</td>\n",
       "      <td>2LD</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.968421</td>\n",
       "      <td>0.967568</td>\n",
       "      <td>0.967742</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578749</th>\n",
       "      <td>1480616853</td>\n",
       "      <td>1029</td>\n",
       "      <td>0</td>\n",
       "      <td>1991</td>\n",
       "      <td>9</td>\n",
       "      <td>False</td>\n",
       "      <td>US</td>\n",
       "      <td>C4R</td>\n",
       "      <td>0.878378</td>\n",
       "      <td>0.868066</td>\n",
       "      <td>0.777778</td>\n",
       "      <td>0.875000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578750</th>\n",
       "      <td>1479644510</td>\n",
       "      <td>4630</td>\n",
       "      <td>0</td>\n",
       "      <td>1991</td>\n",
       "      <td>9</td>\n",
       "      <td>False</td>\n",
       "      <td>US</td>\n",
       "      <td>C4R</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.837838</td>\n",
       "      <td>0.875527</td>\n",
       "      <td>0.864198</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7578751</th>\n",
       "      <td>1480491885</td>\n",
       "      <td>6467</td>\n",
       "      <td>0</td>\n",
       "      <td>1991</td>\n",
       "      <td>9</td>\n",
       "      <td>False</td>\n",
       "      <td>US</td>\n",
       "      <td>C4R</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.913043</td>\n",
       "      <td>0.931751</td>\n",
       "      <td>0.977273</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          ts_listen  user_id  is_listened  release_date_year  \\\n",
       "7578747  1479664431    12641            0               2016   \n",
       "7578748  1480358099    10055            0               2016   \n",
       "7578749  1480616853     1029            0               1991   \n",
       "7578750  1479644510     4630            0               1991   \n",
       "7578751  1480491885     6467            0               1991   \n",
       "\n",
       "         release_date_decade ExplicitLyrics Isrc_Country Isrc_Label  \\\n",
       "7578747                   11          False           GB        2LD   \n",
       "7578748                   11          False           GB        2LD   \n",
       "7578749                    9          False           US        C4R   \n",
       "7578750                    9          False           US        C4R   \n",
       "7578751                    9          False           US        C4R   \n",
       "\n",
       "         release_date_year_cumulative_mean  \\\n",
       "7578747                           0.116667   \n",
       "7578748                           1.000000   \n",
       "7578749                           0.878378   \n",
       "7578750                           0.000000   \n",
       "7578751                           0.000000   \n",
       "\n",
       "         release_date_decade_cumulative_mean  ExplicitLyrics_cumulative_mean  \\\n",
       "7578747                             0.190000                        0.224490   \n",
       "7578748                             0.968421                        0.967568   \n",
       "7578749                             0.868066                        0.777778   \n",
       "7578750                             0.837838                        0.875527   \n",
       "7578751                             0.913043                        0.931751   \n",
       "\n",
       "         Isrc_Country_cumulative_mean  Isrc_Label_cumulative_mean  \n",
       "7578747                      0.288462                         0.0  \n",
       "7578748                      0.967742                         1.0  \n",
       "7578749                      0.875000                         0.0  \n",
       "7578750                      0.864198                         0.0  \n",
       "7578751                      0.977273                         0.0  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sort_index(inplace=True)\n",
    "df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "path_to_test_cum_mean = './data/features/cum_mean2.test.csv'\n",
    "path_to_train_cum_mean = './data/features/cum_mean2.train.csv'\n",
    "\n",
    "test_cum_mean = df.iloc[range(len(train), len(train)+len(test))][new_cols]\n",
    "test_cum_mean.to_csv(path_to_test_cum_mean, encoding='utf-8', index=False)\n",
    "\n",
    "train_cum_mean = df.iloc[range(len(train))][new_cols]\n",
    "train_cum_mean.to_csv(path_to_train_cum_mean, encoding='utf-8', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}