export.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline\n",
    "import pandas as pd\n",
    "import datetime as dt\n",
    "import os\n",
    "# os.chdir('../switch_to_sec/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data = pd.read_csv('../second_data/JData_Action_201604.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data['date'] = data.time.apply(lambda x: x.split(' ')[0])\n",
    "data['hour'] = data.time.apply(lambda x: x.split(' ')[1])\n",
    "del data['time']\n",
    "del data['model_id']\n",
    "del data['brand']\n",
    "data.columns = ['user_id','item_id','behavior_type','cate','brand','date','hour']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "u = data\n",
    "i = pd.read_csv('JData_Product.csv', index_col=['sku_id'])\n",
    "user = pd.read_csv('JData_User.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "105321"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "user.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "24187"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "i.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "u.columns = [u'user_id', u'sku_id', u'type', u'item_category', u'date',\n",
    "       u'hour']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![Paste_Image.png](http://upload-images.jianshu.io/upload_images/1132123-86518acb2e1f623a.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# 每天下单的数量(user, item, ui)\n",
    "df = u[u.type == 4].groupby(['date'])\\\n",
    ".apply(lambda df: pd.Series([\n",
    "            len(df[['user_id', 'sku_id']].drop_duplicates()),\n",
    "            len(df['sku_id'].drop_duplicates()),\n",
    "            len(df['user_id'].drop_duplicates())\n",
    "        ], index=['(u,i) behaviors', 'items', 'users']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>date</th>\n",
       "      <th>2016-04-01</th>\n",
       "      <th>2016-04-02</th>\n",
       "      <th>2016-04-03</th>\n",
       "      <th>2016-04-04</th>\n",
       "      <th>2016-04-05</th>\n",
       "      <th>2016-04-06</th>\n",
       "      <th>2016-04-07</th>\n",
       "      <th>2016-04-08</th>\n",
       "      <th>2016-04-09</th>\n",
       "      <th>2016-04-10</th>\n",
       "      <th>2016-04-11</th>\n",
       "      <th>2016-04-12</th>\n",
       "      <th>2016-04-13</th>\n",
       "      <th>2016-04-14</th>\n",
       "      <th>2016-04-15</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>(u,i) behaviors</th>\n",
       "      <td>207</td>\n",
       "      <td>198</td>\n",
       "      <td>174</td>\n",
       "      <td>175</td>\n",
       "      <td>180</td>\n",
       "      <td>199</td>\n",
       "      <td>445</td>\n",
       "      <td>493</td>\n",
       "      <td>331</td>\n",
       "      <td>249</td>\n",
       "      <td>215</td>\n",
       "      <td>213</td>\n",
       "      <td>203</td>\n",
       "      <td>379</td>\n",
       "      <td>375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>items</th>\n",
       "      <td>105</td>\n",
       "      <td>97</td>\n",
       "      <td>95</td>\n",
       "      <td>99</td>\n",
       "      <td>98</td>\n",
       "      <td>105</td>\n",
       "      <td>142</td>\n",
       "      <td>161</td>\n",
       "      <td>141</td>\n",
       "      <td>112</td>\n",
       "      <td>117</td>\n",
       "      <td>102</td>\n",
       "      <td>98</td>\n",
       "      <td>136</td>\n",
       "      <td>170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>users</th>\n",
       "      <td>204</td>\n",
       "      <td>198</td>\n",
       "      <td>169</td>\n",
       "      <td>172</td>\n",
       "      <td>177</td>\n",
       "      <td>196</td>\n",
       "      <td>441</td>\n",
       "      <td>490</td>\n",
       "      <td>326</td>\n",
       "      <td>249</td>\n",
       "      <td>215</td>\n",
       "      <td>208</td>\n",
       "      <td>203</td>\n",
       "      <td>379</td>\n",
       "      <td>375</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "date             2016-04-01  2016-04-02  2016-04-03  2016-04-04  2016-04-05  \\\n",
       "(u,i) behaviors         207         198         174         175         180   \n",
       "items                   105          97          95          99          98   \n",
       "users                   204         198         169         172         177   \n",
       "\n",
       "date             2016-04-06  2016-04-07  2016-04-08  2016-04-09  2016-04-10  \\\n",
       "(u,i) behaviors         199         445         493         331         249   \n",
       "items                   105         142         161         141         112   \n",
       "users                   196         441         490         326         249   \n",
       "\n",
       "date             2016-04-11  2016-04-12  2016-04-13  2016-04-14  2016-04-15  \n",
       "(u,i) behaviors         215         213         203         379         375  \n",
       "items                   117         102          98         136         170  \n",
       "users                   215         208         203         379         375  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA2oAAAE/CAYAAAA39zBmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4e10e48fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 每个用户对商品子集的购买欲望不强，每个用户平均只会买一个\n",
    "df.plot(kind='bar', figsize=(15, 4));\n",
    "df.T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "u['week'] = u.date.map(lambda d: dt.datetime.strptime(d, '%Y-%m-%d').strftime('%w'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>week</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>(u,i) pairs</th>\n",
       "      <td>73454</td>\n",
       "      <td>71676</td>\n",
       "      <td>66843</td>\n",
       "      <td>66414</td>\n",
       "      <td>83915</td>\n",
       "      <td>120364</td>\n",
       "      <td>80894</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>user</th>\n",
       "      <td>20729</td>\n",
       "      <td>20184</td>\n",
       "      <td>18837</td>\n",
       "      <td>19167</td>\n",
       "      <td>23519</td>\n",
       "      <td>30809</td>\n",
       "      <td>22892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>item</th>\n",
       "      <td>2386</td>\n",
       "      <td>2394</td>\n",
       "      <td>2381</td>\n",
       "      <td>2383</td>\n",
       "      <td>2402</td>\n",
       "      <td>2624</td>\n",
       "      <td>2395</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "week             0      1      2      3      4       5      6\n",
       "(u,i) pairs  73454  71676  66843  66414  83915  120364  80894\n",
       "user         20729  20184  18837  19167  23519   30809  22892\n",
       "item          2386   2394   2381   2383   2402    2624   2395"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmYAAAGFCAYAAABXFD87AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n+3xHiquIAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2a2d660050>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 平均每个用户对很少的商品子集操作\n",
    "df = u[u.type.lt(7) & u.date.lt('2016-04-16')].groupby(['week'])\\\n",
    ".apply(lambda df: pd.Series([\n",
    "            len(df[['user_id', 'sku_id']].drop_duplicates()),\n",
    "            len(df['user_id'].drop_duplicates()),\n",
    "            len(df['sku_id'].drop_duplicates())\n",
    "        ], index=['(u,i) pairs', 'user', 'item']))\n",
    "df.plot(kind='bar', figsize=(10, 6), title=\"04/01~04/15\");\n",
    "df.T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f2a2d67b350>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPwAAAD8CAYAAABTq8lnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n3e2M6HPdhUgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2a4299e890>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 总的点击，浏览，收藏，加购，删除购物车，　下单的比例\n",
    "u[u.date.lt('2016-04-16')].groupby('type').size().plot(kind='pie',autopct='%.2f',figsize=(4, 4),title=\"Distribution of Behavior Types\",legend=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f2a2d493090>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAEVCAYAAAAckrn/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMYeGmZkV+/86tTDrBxvEmwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2a358c8490>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 通常在星期五对商品子集的操作最多\n",
    "u[u.date.lt('2016-04-16')].groupby('week').size().plot(kind='bar', title='Records in Weekday (< 04/16)')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5QAAAE6CAYAAABpvx3/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2a35963e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 总的操作跟下单大致相当\n",
    "pd.DataFrame(dict(\n",
    "    buy=u[u.type.eq(4)].groupby('date').size(),\n",
    "    etc=u[u.type.ne(4)].groupby('date').size()\n",
    ")).plot(kind='bar', secondary_y=['etc'], title='Records in Dates', figsize=(15,4));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![Paste_Image.png](http://upload-images.jianshu.io/upload_images/1132123-ba462757349a7983.png?imageMogr2/auto-orient/strip%7CimageView2/2/w/1240)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>sku_id</th>\n",
       "      <th>type</th>\n",
       "      <th>item_category</th>\n",
       "      <th>date</th>\n",
       "      <th>hour</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>208266</td>\n",
       "      <td>31662</td>\n",
       "      <td>1</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>209390</td>\n",
       "      <td>118799</td>\n",
       "      <td>6</td>\n",
       "      <td>244</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>237311</td>\n",
       "      <td>5825</td>\n",
       "      <td>1</td>\n",
       "      <td>885</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>297325</td>\n",
       "      <td>128747</td>\n",
       "      <td>6</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>287896</td>\n",
       "      <td>68767</td>\n",
       "      <td>6</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:07</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id  sku_id  type  item_category        date      hour\n",
       "0   208266   31662     1            545  2016-03-31  23:59:04\n",
       "1   209390  118799     6            244  2016-03-31  23:59:05\n",
       "2   237311    5825     1            885  2016-03-31  23:59:05\n",
       "3   297325  128747     6            489  2016-03-31  23:59:06\n",
       "4   287896   68767     6            545  2016-03-31  23:59:07"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "u.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "u.columns = ['user_id','item_id','behavior_type','brand','date','hour']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAG0CAYAAADwwWkLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4e1df81f90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 距离考察日之前的ui元组出现在考察日那一天的数量占考察日发生购买行为的ｕｉ的比例为多小\n",
    "sel_ui = u[u.behavior_type.eq(4) & u.date.lt('2016-04-16')&u.date.ge('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id'])\n",
    "pd.DataFrame({\n",
    "    'recent 1 day': [\n",
    "        sum(u[u.date.ge('2016-04-10') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-10') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-10') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui))\n",
    "    ],\n",
    "    'recent 2 days': [\n",
    "        sum(u[u.date.ge('2016-04-09') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-09') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-09') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui))\n",
    "    ],\n",
    "    'recent 3 days': [\n",
    "        sum(u[u.date.ge('2016-04-08') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-08') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-08') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) /float(len(sel_ui))\n",
    "    ],\n",
    "    'recent 4 days': [\n",
    "        sum(u[u.date.ge('2016-04-07') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-07') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-07') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui))\n",
    "    ],\n",
    "    'recent 5 days': [\n",
    "        sum(u[u.date.ge('2016-04-06') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-06') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui)),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-06') & u.date.lt('2016-04-11')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(sel_ui))\n",
    "    ]\n",
    "}, index=['All records', 'Cart records', 'Cart/Del_cart/Buy/focus records']).T.plot(kind='bar', title='Ratio of Matched Records in Selected Date', figsize=(10,6));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAl0AAAG0CAYAAAAIIZL8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2a2d66a450>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 考察日之前浏览，收藏，加购的次数对应的在考察日的浏览，收藏，加购的比例为多小．算式: 考察日之前intersection(在考察日的) / 考察日之前\n",
    "# ex:在考察日之前的10%的加购物车行为，转变成考察日的购买行为\n",
    "sel_ui = u[u.behavior_type.eq(4) & u.date.eq('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id'])\n",
    "pd.DataFrame({\n",
    "    'recent 1 day': [\n",
    "        sum(u[u.date.ge('2016-04-14') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.date.ge('2016-04-14') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-14') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.behavior_type.eq(2) & u.date.ge('2016-04-14') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-14') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-14') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates()))\n",
    "    ],\n",
    "    'recent 2 days': [\n",
    "        sum(u[u.date.ge('2016-04-13') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.date.ge('2016-04-13') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-13') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.behavior_type.eq(2) & u.date.ge('2016-04-13') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-13') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-13') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates()))\n",
    "    ],\n",
    "    'recent 3 days': [\n",
    "        sum(u[u.date.ge('2016-04-12') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.date.ge('2016-04-12') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-12') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.behavior_type.eq(2) & u.date.ge('2016-04-12') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-12') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) /float(len(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-12') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates()))\n",
    "    ],\n",
    "    'recent 4 days': [\n",
    "        sum(u[u.date.ge('2016-04-11') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.date.ge('2016-04-11') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-11') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.behavior_type.eq(2) & u.date.ge('2016-04-11') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-11') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-11') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates()))\n",
    "    ],\n",
    "    'recent 5 days': [\n",
    "        sum(u[u.date.ge('2016-04-10') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.date.ge('2016-04-10') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.eq(2) & u.date.ge('2016-04-10') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.behavior_type.eq(2) & u.date.ge('2016-04-10') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates())),\n",
    "        sum(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-10') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates().set_index(['user_id', 'item_id']).index.isin(sel_ui.index)) / float(len(u[u.behavior_type.lt(6) & u.behavior_type.ge(2) & u.date.ge('2016-04-10') & u.date.lt('2016-04-15')][['user_id', 'item_id']].drop_duplicates()))\n",
    "    ]\n",
    "}, index=['All records', 'Cart records', 'Cart/Del_cart/Buy/focus records']).T.plot(kind='bar', title='Ratio of Matched Records in Selected Date', figsize=(10,6));"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 分析个人数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "udf = u.set_index(['user_id', 'item_id'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "u14 = udf[udf.behavior_type.eq(2) & udf.date.eq('2016-04-14')] # 14号加购物车"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "u15 = udf[udf.behavior_type.eq(4) & udf.date.eq('2016-04-15')] # 15号购买"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# 未来7天\n",
    "udf2 = u[u.date.lt('2016-04-15') & u.date.ge('2016-04-08')].set_index(['user_id', 'brand', 'item_id'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "udf = u.set_index(['user_id', 'brand', 'item_id'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def inspect_user_behavior_percentage(user_id, cate, item_id = 0):\n",
    "    # 考察日之前用户操作商品种类的数量和商品数量\n",
    "    fig, axes = plt.subplots(nrows = 1, ncols = 2, figsize = (12,5))\n",
    "    udf2.loc[user_id, cate, :].sort_values(by=['date', 'hour']).reset_index().groupby('item_id').size().plot(kind = 'pie', ax= axes[0], autopct = '%.1f')\n",
    "    udf2.loc[user_id, :, :].groupby(level = 0).size().plot(kind = 'pie', ax = axes[1], autopct = '%.1f')\n",
    "    return udf2[udf2.behavior_type == 2].loc[user_id, :, :].sort_values(by = ['date', 'hour'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>no</th>\n",
       "      <th>model_id</th>\n",
       "      <th>behavior_type</th>\n",
       "      <th>cate</th>\n",
       "      <th>brand</th>\n",
       "      <th>date</th>\n",
       "      <th>hour</th>\n",
       "      <th>week</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>user_id</th>\n",
       "      <th>item_id</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>77249</th>\n",
       "      <th>154636</th>\n",
       "      <td>18237117</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>16:22:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45064</th>\n",
       "      <th>154636</th>\n",
       "      <td>18241896</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>12:17:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3890</th>\n",
       "      <th>14433</th>\n",
       "      <td>18247818</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>07:20:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57797</th>\n",
       "      <th>14433</th>\n",
       "      <td>18248067</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>14:28:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39350</th>\n",
       "      <th>50482</th>\n",
       "      <td>18255660</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>05:06:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13220</th>\n",
       "      <th>16839</th>\n",
       "      <td>18259184</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>16:03:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70695</th>\n",
       "      <th>125756</th>\n",
       "      <td>18260204</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>21:22:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65360</th>\n",
       "      <th>125756</th>\n",
       "      <td>18262164</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>20:00:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10010</th>\n",
       "      <th>16839</th>\n",
       "      <td>18262215</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>14:50:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6345</th>\n",
       "      <th>14577</th>\n",
       "      <td>18262404</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>677</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>10:28:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84669</th>\n",
       "      <th>62863</th>\n",
       "      <td>18263823</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>17:25:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44489</th>\n",
       "      <th>11090</th>\n",
       "      <td>18264447</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>00:07:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96608</th>\n",
       "      <th>125756</th>\n",
       "      <td>18266336</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>09:18:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87675</th>\n",
       "      <th>131777</th>\n",
       "      <td>18268145</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>677</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>20:45:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27725</th>\n",
       "      <th>144549</th>\n",
       "      <td>18271028</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>16:39:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5670</th>\n",
       "      <th>144549</th>\n",
       "      <td>18271187</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>14:33:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11152</th>\n",
       "      <th>117452</th>\n",
       "      <td>18274219</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>677</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>14:52:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18610</th>\n",
       "      <th>20878</th>\n",
       "      <td>18284484</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>13:30:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44191</th>\n",
       "      <th>14181</th>\n",
       "      <td>18286292</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>10:25:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42568</th>\n",
       "      <th>14181</th>\n",
       "      <td>18287214</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>21:02:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15581</th>\n",
       "      <th>37957</th>\n",
       "      <td>18288324</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>09:58:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5727</th>\n",
       "      <th>95850</th>\n",
       "      <td>18290626</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>800</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>23:33:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17277</th>\n",
       "      <th>37957</th>\n",
       "      <td>18292531</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>16:56:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51450</th>\n",
       "      <th>65233</th>\n",
       "      <td>18297883</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>800</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>16:45:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43343</th>\n",
       "      <th>65233</th>\n",
       "      <td>18297938</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>800</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>12:14:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3636</th>\n",
       "      <th>116489</th>\n",
       "      <td>18301124</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>21:49:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4271</th>\n",
       "      <th>116489</th>\n",
       "      <td>18302481</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>10:59:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78524</th>\n",
       "      <th>84389</th>\n",
       "      <td>18306811</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>23:15:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12490</th>\n",
       "      <th>105064</th>\n",
       "      <td>18314286</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>00:01:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39878</th>\n",
       "      <th>106714</th>\n",
       "      <td>18318818</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>00:10:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20486</th>\n",
       "      <th>38594</th>\n",
       "      <td>18667819</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>244</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>15:34:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47975</th>\n",
       "      <th>21147</th>\n",
       "      <td>18668046</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>403</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>13:25:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51758</th>\n",
       "      <th>21147</th>\n",
       "      <td>18668878</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>403</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>16:10:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88287</th>\n",
       "      <th>57018</th>\n",
       "      <td>18674134</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>09:33:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55197</th>\n",
       "      <th>133192</th>\n",
       "      <td>18677616</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>800</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>21:45:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6042</th>\n",
       "      <th>57018</th>\n",
       "      <td>18679008</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>13:49:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9499</th>\n",
       "      <th>57018</th>\n",
       "      <td>18682713</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>17:39:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86944</th>\n",
       "      <th>70839</th>\n",
       "      <td>18684589</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>16:13:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86636</th>\n",
       "      <th>24679</th>\n",
       "      <td>18686926</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>658</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>11:59:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3781</th>\n",
       "      <th>41101</th>\n",
       "      <td>18689876</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>658</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>12:43:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16342</th>\n",
       "      <th>41101</th>\n",
       "      <td>18690096</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>658</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>16:39:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90597</th>\n",
       "      <th>57018</th>\n",
       "      <td>18699533</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>15:17:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65126</th>\n",
       "      <th>57018</th>\n",
       "      <td>18704212</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>13:57:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64903</th>\n",
       "      <th>57018</th>\n",
       "      <td>18710411</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>12:05:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62906</th>\n",
       "      <th>103234</th>\n",
       "      <td>18711309</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>18:50:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102255</th>\n",
       "      <th>154636</th>\n",
       "      <td>18715203</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>17:46:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11309</th>\n",
       "      <th>154636</th>\n",
       "      <td>18715250</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>16:55:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103508</th>\n",
       "      <th>21147</th>\n",
       "      <td>18716429</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>403</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>00:42:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63550</th>\n",
       "      <th>154636</th>\n",
       "      <td>18718037</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>17:41:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74551</th>\n",
       "      <th>154636</th>\n",
       "      <td>18718510</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>21:47:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30080</th>\n",
       "      <th>154636</th>\n",
       "      <td>18720676</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>21:32:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57966</th>\n",
       "      <th>154636</th>\n",
       "      <td>18722169</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>10:50:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76839</th>\n",
       "      <th>154636</th>\n",
       "      <td>18724540</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>22:08:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48035</th>\n",
       "      <th>154636</th>\n",
       "      <td>18724541</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>19:15:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53043</th>\n",
       "      <th>154636</th>\n",
       "      <td>18728796</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>10:48:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12214</th>\n",
       "      <th>154636</th>\n",
       "      <td>18730126</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>14:00:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79229</th>\n",
       "      <th>49269</th>\n",
       "      <td>18732604</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>857</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>10:28:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3151</th>\n",
       "      <th>154636</th>\n",
       "      <td>18734984</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>00:39:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21833</th>\n",
       "      <th>154636</th>\n",
       "      <td>18737619</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>23:12:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77603</th>\n",
       "      <th>154636</th>\n",
       "      <td>18741477</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-15</td>\n",
       "      <td>10:54:00</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>272 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                       no  model_id  behavior_type  cate  brand        date  \\\n",
       "user_id item_id                                                               \n",
       "77249   154636   18237117       NaN              4     8    545  2016-04-15   \n",
       "45064   154636   18241896       NaN              4     8    545  2016-04-15   \n",
       "3890    14433    18247818       NaN              4     8     30  2016-04-15   \n",
       "57797   14433    18248067       NaN              4     8     30  2016-04-15   \n",
       "39350   50482    18255660       NaN              4     8    489  2016-04-15   \n",
       "13220   16839    18259184       NaN              4     8    214  2016-04-15   \n",
       "70695   125756   18260204       NaN              4     8     30  2016-04-15   \n",
       "65360   125756   18262164       NaN              4     8     30  2016-04-15   \n",
       "10010   16839    18262215       NaN              4     8    214  2016-04-15   \n",
       "6345    14577    18262404       NaN              4     8    677  2016-04-15   \n",
       "84669   62863    18263823       NaN              4     8    489  2016-04-15   \n",
       "44489   11090    18264447       NaN              4     8    214  2016-04-15   \n",
       "96608   125756   18266336       NaN              4     8     30  2016-04-15   \n",
       "87675   131777   18268145       NaN              4     8    677  2016-04-15   \n",
       "27725   144549   18271028       NaN              4     8    545  2016-04-15   \n",
       "5670    144549   18271187       NaN              4     8    545  2016-04-15   \n",
       "11152   117452   18274219       NaN              4     8    677  2016-04-15   \n",
       "18610   20878    18284484       NaN              4     8    545  2016-04-15   \n",
       "44191   14181    18286292       NaN              4     8    214  2016-04-15   \n",
       "42568   14181    18287214       NaN              4     8    214  2016-04-15   \n",
       "15581   37957    18288324       NaN              4     8    489  2016-04-15   \n",
       "5727    95850    18290626       NaN              4     8    800  2016-04-15   \n",
       "17277   37957    18292531       NaN              4     8    489  2016-04-15   \n",
       "51450   65233    18297883       NaN              4     8    800  2016-04-15   \n",
       "43343   65233    18297938       NaN              4     8    800  2016-04-15   \n",
       "3636    116489   18301124       NaN              4     8    545  2016-04-15   \n",
       "4271    116489   18302481       NaN              4     8    545  2016-04-15   \n",
       "78524   84389    18306811       NaN              4     8    489  2016-04-15   \n",
       "12490   105064   18314286       NaN              4     8    214  2016-04-15   \n",
       "39878   106714   18318818       NaN              4     8    214  2016-04-15   \n",
       "...                   ...       ...            ...   ...    ...         ...   \n",
       "20486   38594    18667819       NaN              4     8    244  2016-04-15   \n",
       "47975   21147    18668046       NaN              4     8    403  2016-04-15   \n",
       "51758   21147    18668878       NaN              4     8    403  2016-04-15   \n",
       "88287   57018    18674134       NaN              4     8    545  2016-04-15   \n",
       "55197   133192   18677616       NaN              4     8    800  2016-04-15   \n",
       "6042    57018    18679008       NaN              4     8    545  2016-04-15   \n",
       "9499    57018    18682713       NaN              4     8    545  2016-04-15   \n",
       "86944   70839    18684589       NaN              4     8    545  2016-04-15   \n",
       "86636   24679    18686926       NaN              4     8    658  2016-04-15   \n",
       "3781    41101    18689876       NaN              4     8    658  2016-04-15   \n",
       "16342   41101    18690096       NaN              4     8    658  2016-04-15   \n",
       "90597   57018    18699533       NaN              4     8    545  2016-04-15   \n",
       "65126   57018    18704212       NaN              4     8    545  2016-04-15   \n",
       "64903   57018    18710411       NaN              4     8    545  2016-04-15   \n",
       "62906   103234   18711309       NaN              4     8    214  2016-04-15   \n",
       "102255  154636   18715203       NaN              4     8    545  2016-04-15   \n",
       "11309   154636   18715250       NaN              4     8    545  2016-04-15   \n",
       "103508  21147    18716429       NaN              4     8    403  2016-04-15   \n",
       "63550   154636   18718037       NaN              4     8    545  2016-04-15   \n",
       "74551   154636   18718510       NaN              4     8    545  2016-04-15   \n",
       "30080   154636   18720676       NaN              4     8    545  2016-04-15   \n",
       "57966   154636   18722169       NaN              4     8    545  2016-04-15   \n",
       "76839   154636   18724540       NaN              4     8    545  2016-04-15   \n",
       "48035   154636   18724541       NaN              4     8    545  2016-04-15   \n",
       "53043   154636   18728796       NaN              4     8    545  2016-04-15   \n",
       "12214   154636   18730126       NaN              4     8    545  2016-04-15   \n",
       "79229   49269    18732604       NaN              4     8    857  2016-04-15   \n",
       "3151    154636   18734984       NaN              4     8    545  2016-04-15   \n",
       "21833   154636   18737619       NaN              4     8    545  2016-04-15   \n",
       "77603   154636   18741477       NaN              4     8    545  2016-04-15   \n",
       "\n",
       "                     hour week  \n",
       "user_id item_id                 \n",
       "77249   154636   16:22:00    5  \n",
       "45064   154636   12:17:00    5  \n",
       "3890    14433    07:20:00    5  \n",
       "57797   14433    14:28:00    5  \n",
       "39350   50482    05:06:00    5  \n",
       "13220   16839    16:03:00    5  \n",
       "70695   125756   21:22:00    5  \n",
       "65360   125756   20:00:00    5  \n",
       "10010   16839    14:50:00    5  \n",
       "6345    14577    10:28:00    5  \n",
       "84669   62863    17:25:00    5  \n",
       "44489   11090    00:07:00    5  \n",
       "96608   125756   09:18:00    5  \n",
       "87675   131777   20:45:00    5  \n",
       "27725   144549   16:39:00    5  \n",
       "5670    144549   14:33:00    5  \n",
       "11152   117452   14:52:00    5  \n",
       "18610   20878    13:30:00    5  \n",
       "44191   14181    10:25:00    5  \n",
       "42568   14181    21:02:00    5  \n",
       "15581   37957    09:58:00    5  \n",
       "5727    95850    23:33:00    5  \n",
       "17277   37957    16:56:00    5  \n",
       "51450   65233    16:45:00    5  \n",
       "43343   65233    12:14:00    5  \n",
       "3636    116489   21:49:00    5  \n",
       "4271    116489   10:59:00    5  \n",
       "78524   84389    23:15:00    5  \n",
       "12490   105064   00:01:00    5  \n",
       "39878   106714   00:10:00    5  \n",
       "...                   ...  ...  \n",
       "20486   38594    15:34:00    5  \n",
       "47975   21147    13:25:00    5  \n",
       "51758   21147    16:10:00    5  \n",
       "88287   57018    09:33:00    5  \n",
       "55197   133192   21:45:00    5  \n",
       "6042    57018    13:49:00    5  \n",
       "9499    57018    17:39:00    5  \n",
       "86944   70839    16:13:00    5  \n",
       "86636   24679    11:59:00    5  \n",
       "3781    41101    12:43:00    5  \n",
       "16342   41101    16:39:00    5  \n",
       "90597   57018    15:17:00    5  \n",
       "65126   57018    13:57:00    5  \n",
       "64903   57018    12:05:00    5  \n",
       "62906   103234   18:50:00    5  \n",
       "102255  154636   17:46:00    5  \n",
       "11309   154636   16:55:00    5  \n",
       "103508  21147    00:42:00    5  \n",
       "63550   154636   17:41:00    5  \n",
       "74551   154636   21:47:00    5  \n",
       "30080   154636   21:32:00    5  \n",
       "57966   154636   10:50:00    5  \n",
       "76839   154636   22:08:00    5  \n",
       "48035   154636   19:15:00    5  \n",
       "53043   154636   10:48:00    5  \n",
       "12214   154636   14:00:00    5  \n",
       "79229   49269    10:28:00    5  \n",
       "3151    154636   00:39:00    5  \n",
       "21833   154636   23:12:00    5  \n",
       "77603   154636   10:54:00    5  \n",
       "\n",
       "[272 rows x 8 columns]"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 历史购物车是否有此类商品\n",
    "u15[~u15.index.isin(u14.index)] # 除了14号加购物车，15号购买之外的在15号人"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>no</th>\n",
       "      <th>model_id</th>\n",
       "      <th>behavior_type</th>\n",
       "      <th>cate</th>\n",
       "      <th>date</th>\n",
       "      <th>hour</th>\n",
       "      <th>week</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>brand</th>\n",
       "      <th>item_id</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>545</th>\n",
       "      <th>42199</th>\n",
       "      <td>15493443</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>2016-04-13</td>\n",
       "      <td>17:40:22</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <th>14433</th>\n",
       "      <td>15405565</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>2016-04-13</td>\n",
       "      <td>17:41:28</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>545</th>\n",
       "      <th>154636</th>\n",
       "      <td>15558689</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>2016-04-13</td>\n",
       "      <td>17:43:15</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     no  model_id  behavior_type  cate        date      hour  \\\n",
       "brand item_id                                                                  \n",
       "545   42199    15493443       NaN              2     8  2016-04-13  17:40:22   \n",
       "30    14433    15405565       NaN              2     8  2016-04-13  17:41:28   \n",
       "545   154636   15558689       NaN              2     8  2016-04-13  17:43:15   \n",
       "\n",
       "              week  \n",
       "brand item_id       \n",
       "545   42199      3  \n",
       "30    14433      3  \n",
       "545   154636     3  "
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAroAAAElCAYAAAALJUESAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefbe2f0c10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "inspect_user_behavior_percentage(3151, 545, 154636)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>no</th>\n",
       "      <th>model_id</th>\n",
       "      <th>behavior_type</th>\n",
       "      <th>cate</th>\n",
       "      <th>brand</th>\n",
       "      <th>date</th>\n",
       "      <th>hour</th>\n",
       "      <th>week</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>user_id</th>\n",
       "      <th>item_id</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>52919</th>\n",
       "      <th>166707</th>\n",
       "      <td>16562067</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>21:11:21</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60244</th>\n",
       "      <th>64794</th>\n",
       "      <td>16564363</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>23:27:51</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63645</th>\n",
       "      <th>151327</th>\n",
       "      <td>16582605</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>403</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>21:01:46</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92794</th>\n",
       "      <th>149641</th>\n",
       "      <td>16606067</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>403</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>21:18:24</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">34266</th>\n",
       "      <th>79636</th>\n",
       "      <td>16625189</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>22:05:41</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79636</th>\n",
       "      <td>16625190</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>22:04:29</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49135</th>\n",
       "      <th>78694</th>\n",
       "      <td>16626547</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>244</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>12:35:31</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74231</th>\n",
       "      <th>84389</th>\n",
       "      <td>16628975</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>16:28:19</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58721</th>\n",
       "      <th>84389</th>\n",
       "      <td>16632063</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>20:40:07</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6520</th>\n",
       "      <th>145946</th>\n",
       "      <td>16712034</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>658</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>20:11:06</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15883</th>\n",
       "      <th>145946</th>\n",
       "      <td>16712137</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>658</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>22:04:39</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8804</th>\n",
       "      <th>69355</th>\n",
       "      <td>16764033</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>800</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>17:55:10</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59418</th>\n",
       "      <th>37957</th>\n",
       "      <td>16767441</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>14:40:33</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33619</th>\n",
       "      <th>1628</th>\n",
       "      <td>16768456</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>20:06:46</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57019</th>\n",
       "      <th>11090</th>\n",
       "      <td>16772553</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>23:50:00</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89709</th>\n",
       "      <th>117452</th>\n",
       "      <td>16794258</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>677</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>15:55:13</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20962</th>\n",
       "      <th>117452</th>\n",
       "      <td>16795788</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>677</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>13:55:20</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60721</th>\n",
       "      <th>116489</th>\n",
       "      <td>16797587</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>17:21:48</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">9653</th>\n",
       "      <th>158013</th>\n",
       "      <td>16806034</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>16:53:04</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>158013</th>\n",
       "      <td>16806041</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>10:47:33</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84585</th>\n",
       "      <th>12540</th>\n",
       "      <td>16809220</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>22:21:04</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79526</th>\n",
       "      <th>103014</th>\n",
       "      <td>16833860</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>800</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>21:16:14</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">100786</th>\n",
       "      <th>131300</th>\n",
       "      <td>16843146</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>693</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>04:07:32</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131300</th>\n",
       "      <td>16843148</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>693</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>04:07:45</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">46877</th>\n",
       "      <th>138151</th>\n",
       "      <td>16860308</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>10:45:04</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>138151</th>\n",
       "      <td>16860309</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>10:43:17</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44604</th>\n",
       "      <th>138151</th>\n",
       "      <td>16861505</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>30</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>08:43:14</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71669</th>\n",
       "      <th>36748</th>\n",
       "      <td>16865476</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>427</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>13:23:02</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33588</th>\n",
       "      <th>115867</th>\n",
       "      <td>16871005</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>08:44:06</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93795</th>\n",
       "      <th>62351</th>\n",
       "      <td>16879390</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>306</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>22:38:50</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">12806</th>\n",
       "      <th>160476</th>\n",
       "      <td>17056549</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>12:42:16</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>160476</th>\n",
       "      <td>17056550</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>12:46:11</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>160476</th>\n",
       "      <td>17056552</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>21:34:14</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31082</th>\n",
       "      <th>63006</th>\n",
       "      <td>17081579</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>09:09:55</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">38761</th>\n",
       "      <th>61531</th>\n",
       "      <td>17081656</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>09:06:28</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61531</th>\n",
       "      <td>17081657</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>21:53:39</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">95452</th>\n",
       "      <th>12564</th>\n",
       "      <td>17085016</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>11:55:12</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12564</th>\n",
       "      <td>17085017</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>11:56:03</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12564</th>\n",
       "      <td>17085018</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>11:54:14</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12564</th>\n",
       "      <td>17085019</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>11:51:47</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58574</th>\n",
       "      <th>89802</th>\n",
       "      <td>17085801</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>11:18:41</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33436</th>\n",
       "      <th>21147</th>\n",
       "      <td>17089218</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>403</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>16:18:56</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">2415</th>\n",
       "      <th>12564</th>\n",
       "      <td>17091085</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>16:51:38</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12564</th>\n",
       "      <td>17091086</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>16:48:11</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12564</th>\n",
       "      <td>17091087</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>16:50:09</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>83313</th>\n",
       "      <th>18412</th>\n",
       "      <td>17098660</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>15:07:42</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59597</th>\n",
       "      <th>152872</th>\n",
       "      <td>17116720</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>306</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>00:29:44</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30522</th>\n",
       "      <th>128747</th>\n",
       "      <td>17130916</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>22:41:19</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6370</th>\n",
       "      <th>32465</th>\n",
       "      <td>17135810</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>18:46:38</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">12135</th>\n",
       "      <th>127735</th>\n",
       "      <td>17140435</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>14:16:58</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127735</th>\n",
       "      <td>17140436</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>19:56:54</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127735</th>\n",
       "      <td>17140437</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>19:53:49</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86521</th>\n",
       "      <th>7746</th>\n",
       "      <td>17142899</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>22:15:13</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78684</th>\n",
       "      <th>65700</th>\n",
       "      <td>17144473</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>14:29:14</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72854</th>\n",
       "      <th>119111</th>\n",
       "      <td>17145425</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>214</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>13:17:00</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">24856</th>\n",
       "      <th>52343</th>\n",
       "      <td>17148116</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>10:13:42</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52343</th>\n",
       "      <td>17148119</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>09:30:21</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52343</th>\n",
       "      <td>17148121</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>10:13:40</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">97207</th>\n",
       "      <th>56792</th>\n",
       "      <td>17166381</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>18:34:26</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56792</th>\n",
       "      <td>17166382</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>18:11:55</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>81 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                       no  model_id  behavior_type  cate  brand        date  \\\n",
       "user_id item_id                                                               \n",
       "52919   166707   16562067       NaN              2     8    489  2016-04-14   \n",
       "60244   64794    16564363       NaN              2     8     30  2016-04-14   \n",
       "63645   151327   16582605       NaN              2     8    403  2016-04-14   \n",
       "92794   149641   16606067       NaN              2     8    403  2016-04-14   \n",
       "34266   79636    16625189       NaN              2     8    489  2016-04-14   \n",
       "        79636    16625190       NaN              2     8    489  2016-04-14   \n",
       "49135   78694    16626547       NaN              2     8    244  2016-04-14   \n",
       "74231   84389    16628975       NaN              2     8    489  2016-04-14   \n",
       "58721   84389    16632063       NaN              2     8    489  2016-04-14   \n",
       "6520    145946   16712034       NaN              2     8    658  2016-04-14   \n",
       "15883   145946   16712137       NaN              2     8    658  2016-04-14   \n",
       "8804    69355    16764033       NaN              2     8    800  2016-04-14   \n",
       "59418   37957    16767441       NaN              2     8    489  2016-04-14   \n",
       "33619   1628     16768456       NaN              2     8    214  2016-04-14   \n",
       "57019   11090    16772553       NaN              2     8    214  2016-04-14   \n",
       "89709   117452   16794258       NaN              2     8    677  2016-04-14   \n",
       "20962   117452   16795788       NaN              2     8    677  2016-04-14   \n",
       "60721   116489   16797587       NaN              2     8    545  2016-04-14   \n",
       "9653    158013   16806034       NaN              2     8    214  2016-04-14   \n",
       "        158013   16806041       NaN              2     8    214  2016-04-14   \n",
       "84585   12540    16809220       NaN              2     8    545  2016-04-14   \n",
       "79526   103014   16833860       NaN              2     8    800  2016-04-14   \n",
       "100786  131300   16843146       NaN              2     8    693  2016-04-14   \n",
       "        131300   16843148       NaN              2     8    693  2016-04-14   \n",
       "46877   138151   16860308       NaN              2     8     30  2016-04-14   \n",
       "        138151   16860309       NaN              2     8     30  2016-04-14   \n",
       "44604   138151   16861505       NaN              2     8     30  2016-04-14   \n",
       "71669   36748    16865476       NaN              2     8    427  2016-04-14   \n",
       "33588   115867   16871005       NaN              2     8    489  2016-04-14   \n",
       "93795   62351    16879390       NaN              2     8    306  2016-04-14   \n",
       "...                   ...       ...            ...   ...    ...         ...   \n",
       "12806   160476   17056549       NaN              2     8    545  2016-04-14   \n",
       "        160476   17056550       NaN              2     8    545  2016-04-14   \n",
       "        160476   17056552       NaN              2     8    545  2016-04-14   \n",
       "31082   63006    17081579       NaN              2     8    214  2016-04-14   \n",
       "38761   61531    17081656       NaN              2     8    214  2016-04-14   \n",
       "        61531    17081657       NaN              2     8    214  2016-04-14   \n",
       "95452   12564    17085016       NaN              2     8    214  2016-04-14   \n",
       "        12564    17085017       NaN              2     8    214  2016-04-14   \n",
       "        12564    17085018       NaN              2     8    214  2016-04-14   \n",
       "        12564    17085019       NaN              2     8    214  2016-04-14   \n",
       "58574   89802    17085801       NaN              2     8    545  2016-04-14   \n",
       "33436   21147    17089218       NaN              2     8    403  2016-04-14   \n",
       "2415    12564    17091085       NaN              2     8    214  2016-04-14   \n",
       "        12564    17091086       NaN              2     8    214  2016-04-14   \n",
       "        12564    17091087       NaN              2     8    214  2016-04-14   \n",
       "83313   18412    17098660       NaN              2     8    214  2016-04-14   \n",
       "59597   152872   17116720       NaN              2     8    306  2016-04-14   \n",
       "30522   128747   17130916       NaN              2     8    489  2016-04-14   \n",
       "6370    32465    17135810       NaN              2     8    489  2016-04-14   \n",
       "12135   127735   17140435       NaN              2     8    214  2016-04-14   \n",
       "        127735   17140436       NaN              2     8    214  2016-04-14   \n",
       "        127735   17140437       NaN              2     8    214  2016-04-14   \n",
       "86521   7746     17142899       NaN              2     8    214  2016-04-14   \n",
       "78684   65700    17144473       NaN              2     8    214  2016-04-14   \n",
       "72854   119111   17145425       NaN              2     8    214  2016-04-14   \n",
       "24856   52343    17148116       NaN              2     8    489  2016-04-14   \n",
       "        52343    17148119       NaN              2     8    489  2016-04-14   \n",
       "        52343    17148121       NaN              2     8    489  2016-04-14   \n",
       "97207   56792    17166381       NaN              2     8    489  2016-04-14   \n",
       "        56792    17166382       NaN              2     8    489  2016-04-14   \n",
       "\n",
       "                     hour week  \n",
       "user_id item_id                 \n",
       "52919   166707   21:11:21    4  \n",
       "60244   64794    23:27:51    4  \n",
       "63645   151327   21:01:46    4  \n",
       "92794   149641   21:18:24    4  \n",
       "34266   79636    22:05:41    4  \n",
       "        79636    22:04:29    4  \n",
       "49135   78694    12:35:31    4  \n",
       "74231   84389    16:28:19    4  \n",
       "58721   84389    20:40:07    4  \n",
       "6520    145946   20:11:06    4  \n",
       "15883   145946   22:04:39    4  \n",
       "8804    69355    17:55:10    4  \n",
       "59418   37957    14:40:33    4  \n",
       "33619   1628     20:06:46    4  \n",
       "57019   11090    23:50:00    4  \n",
       "89709   117452   15:55:13    4  \n",
       "20962   117452   13:55:20    4  \n",
       "60721   116489   17:21:48    4  \n",
       "9653    158013   16:53:04    4  \n",
       "        158013   10:47:33    4  \n",
       "84585   12540    22:21:04    4  \n",
       "79526   103014   21:16:14    4  \n",
       "100786  131300   04:07:32    4  \n",
       "        131300   04:07:45    4  \n",
       "46877   138151   10:45:04    4  \n",
       "        138151   10:43:17    4  \n",
       "44604   138151   08:43:14    4  \n",
       "71669   36748    13:23:02    4  \n",
       "33588   115867   08:44:06    4  \n",
       "93795   62351    22:38:50    4  \n",
       "...                   ...  ...  \n",
       "12806   160476   12:42:16    4  \n",
       "        160476   12:46:11    4  \n",
       "        160476   21:34:14    4  \n",
       "31082   63006    09:09:55    4  \n",
       "38761   61531    09:06:28    4  \n",
       "        61531    21:53:39    4  \n",
       "95452   12564    11:55:12    4  \n",
       "        12564    11:56:03    4  \n",
       "        12564    11:54:14    4  \n",
       "        12564    11:51:47    4  \n",
       "58574   89802    11:18:41    4  \n",
       "33436   21147    16:18:56    4  \n",
       "2415    12564    16:51:38    4  \n",
       "        12564    16:48:11    4  \n",
       "        12564    16:50:09    4  \n",
       "83313   18412    15:07:42    4  \n",
       "59597   152872   00:29:44    4  \n",
       "30522   128747   22:41:19    4  \n",
       "6370    32465    18:46:38    4  \n",
       "12135   127735   14:16:58    4  \n",
       "        127735   19:56:54    4  \n",
       "        127735   19:53:49    4  \n",
       "86521   7746     22:15:13    4  \n",
       "78684   65700    14:29:14    4  \n",
       "72854   119111   13:17:00    4  \n",
       "24856   52343    10:13:42    4  \n",
       "        52343    09:30:21    4  \n",
       "        52343    10:13:40    4  \n",
       "97207   56792    18:34:26    4  \n",
       "        56792    18:11:55    4  \n",
       "\n",
       "[81 rows x 8 columns]"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "u14[u14.index.isin(u15.index)] # 14号加购物车，15号购买u的"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>no</th>\n",
       "      <th>model_id</th>\n",
       "      <th>behavior_type</th>\n",
       "      <th>cate</th>\n",
       "      <th>date</th>\n",
       "      <th>hour</th>\n",
       "      <th>week</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>brand</th>\n",
       "      <th>item_id</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\">489</th>\n",
       "      <th>37957</th>\n",
       "      <td>11728960</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>2016-04-10</td>\n",
       "      <td>17:28:06</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37957</th>\n",
       "      <td>16767441</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>2016-04-14</td>\n",
       "      <td>14:40:33</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     no  model_id  behavior_type  cate        date      hour  \\\n",
       "brand item_id                                                                  \n",
       "489   37957    11728960       NaN              2     8  2016-04-10  17:28:06   \n",
       "      37957    16767441       NaN              2     8  2016-04-14  14:40:33   \n",
       "\n",
       "              week  \n",
       "brand item_id       \n",
       "489   37957      0  \n",
       "      37957      4  "
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAArsAAAElCAYAAADk5yosAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fefa9c2f950>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 当天加购，对后一天的影响\n",
    "inspect_user_behavior_percentage(59418, 489, 37957)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "d = u[u.date.lt('2016-04-15') & u.date.ge('2016-04-10')] # 14, 13, 12, 11,10 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "user_last_cart_hour = d[d.behavior_type == 2].groupby('user_id').date.last()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "user_cnt = d.groupby('user_id').behavior_type.sum().reset_index()\n",
    "user_cnt.columns = pd.Index(['user_id', 'user_cnt'])\n",
    "\n",
    "user_item_cnt = d.groupby(['user_id', 'item_id']).agg(dict(behavior_type='sum', brand='first')).reset_index()\n",
    "user_item_cnt.columns = pd.Index(['user_id', 'item_id', 'item_cnt', 'item_category'])\n",
    "\n",
    "user_category_cnt = d.groupby(['user_id', 'brand']).behavior_type.sum().reset_index()\n",
    "user_category_cnt.columns = pd.Index(['user_id', 'item_category', 'category_cnt'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "usm = pd.merge(\n",
    "    user_cnt,\n",
    "    pd.merge(\n",
    "        user_item_cnt,\n",
    "        user_category_cnt,\n",
    "        on=['user_id', 'item_category'], how='left'\n",
    "    ),\n",
    "    on='user_id'\n",
    ")\n",
    "usm['category_user_pcn'] = usm['category_cnt'] / usm['user_cnt']\n",
    "usm['item_category_pcn'] = usm['item_cnt'] / usm['category_cnt']\n",
    "usm['item_user_pcn'] = usm['item_cnt'] / usm['user_cnt']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>user_cnt</th>\n",
       "      <th>item_id</th>\n",
       "      <th>item_cnt</th>\n",
       "      <th>item_category</th>\n",
       "      <th>category_cnt</th>\n",
       "      <th>category_user_pcn</th>\n",
       "      <th>item_category_pcn</th>\n",
       "      <th>item_user_pcn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>76</td>\n",
       "      <td>124997</td>\n",
       "      <td>76</td>\n",
       "      <td>693</td>\n",
       "      <td>76</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>8</td>\n",
       "      <td>6</td>\n",
       "      <td>96887</td>\n",
       "      <td>6</td>\n",
       "      <td>214</td>\n",
       "      <td>6</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>11</td>\n",
       "      <td>90</td>\n",
       "      <td>52343</td>\n",
       "      <td>32</td>\n",
       "      <td>489</td>\n",
       "      <td>32</td>\n",
       "      <td>0.355556</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.355556</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>11</td>\n",
       "      <td>90</td>\n",
       "      <td>57018</td>\n",
       "      <td>58</td>\n",
       "      <td>545</td>\n",
       "      <td>58</td>\n",
       "      <td>0.644444</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.644444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15</td>\n",
       "      <td>90</td>\n",
       "      <td>18412</td>\n",
       "      <td>90</td>\n",
       "      <td>214</td>\n",
       "      <td>90</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>17</td>\n",
       "      <td>109</td>\n",
       "      <td>285</td>\n",
       "      <td>13</td>\n",
       "      <td>214</td>\n",
       "      <td>71</td>\n",
       "      <td>0.651376</td>\n",
       "      <td>0.183099</td>\n",
       "      <td>0.119266</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>17</td>\n",
       "      <td>109</td>\n",
       "      <td>21457</td>\n",
       "      <td>13</td>\n",
       "      <td>214</td>\n",
       "      <td>71</td>\n",
       "      <td>0.651376</td>\n",
       "      <td>0.183099</td>\n",
       "      <td>0.119266</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>17</td>\n",
       "      <td>109</td>\n",
       "      <td>58959</td>\n",
       "      <td>38</td>\n",
       "      <td>622</td>\n",
       "      <td>38</td>\n",
       "      <td>0.348624</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.348624</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>17</td>\n",
       "      <td>109</td>\n",
       "      <td>60077</td>\n",
       "      <td>13</td>\n",
       "      <td>214</td>\n",
       "      <td>71</td>\n",
       "      <td>0.651376</td>\n",
       "      <td>0.183099</td>\n",
       "      <td>0.119266</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>17</td>\n",
       "      <td>109</td>\n",
       "      <td>65428</td>\n",
       "      <td>13</td>\n",
       "      <td>214</td>\n",
       "      <td>71</td>\n",
       "      <td>0.651376</td>\n",
       "      <td>0.183099</td>\n",
       "      <td>0.119266</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>17</td>\n",
       "      <td>109</td>\n",
       "      <td>107318</td>\n",
       "      <td>13</td>\n",
       "      <td>214</td>\n",
       "      <td>71</td>\n",
       "      <td>0.651376</td>\n",
       "      <td>0.183099</td>\n",
       "      <td>0.119266</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>17</td>\n",
       "      <td>109</td>\n",
       "      <td>151195</td>\n",
       "      <td>6</td>\n",
       "      <td>214</td>\n",
       "      <td>71</td>\n",
       "      <td>0.651376</td>\n",
       "      <td>0.084507</td>\n",
       "      <td>0.055046</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>15259</td>\n",
       "      <td>188</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.130737</td>\n",
       "      <td>0.117794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>24371</td>\n",
       "      <td>36</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.025035</td>\n",
       "      <td>0.022556</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>24423</td>\n",
       "      <td>26</td>\n",
       "      <td>214</td>\n",
       "      <td>124</td>\n",
       "      <td>0.077694</td>\n",
       "      <td>0.209677</td>\n",
       "      <td>0.016291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>47950</td>\n",
       "      <td>56</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.038943</td>\n",
       "      <td>0.035088</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>48154</td>\n",
       "      <td>44</td>\n",
       "      <td>214</td>\n",
       "      <td>124</td>\n",
       "      <td>0.077694</td>\n",
       "      <td>0.354839</td>\n",
       "      <td>0.027569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>48895</td>\n",
       "      <td>34</td>\n",
       "      <td>489</td>\n",
       "      <td>34</td>\n",
       "      <td>0.021303</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.021303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>49178</td>\n",
       "      <td>124</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.086231</td>\n",
       "      <td>0.077694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>56234</td>\n",
       "      <td>38</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.026426</td>\n",
       "      <td>0.023810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>75836</td>\n",
       "      <td>28</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.019471</td>\n",
       "      <td>0.017544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>78720</td>\n",
       "      <td>64</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.044506</td>\n",
       "      <td>0.040100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>89802</td>\n",
       "      <td>38</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.026426</td>\n",
       "      <td>0.023810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>108824</td>\n",
       "      <td>66</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.045897</td>\n",
       "      <td>0.041353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>110779</td>\n",
       "      <td>52</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.036161</td>\n",
       "      <td>0.032581</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>116489</td>\n",
       "      <td>26</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.018081</td>\n",
       "      <td>0.016291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>128988</td>\n",
       "      <td>66</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.045897</td>\n",
       "      <td>0.041353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>132562</td>\n",
       "      <td>52</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.036161</td>\n",
       "      <td>0.032581</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>146634</td>\n",
       "      <td>26</td>\n",
       "      <td>545</td>\n",
       "      <td>1438</td>\n",
       "      <td>0.901003</td>\n",
       "      <td>0.018081</td>\n",
       "      <td>0.016291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>22</td>\n",
       "      <td>1596</td>\n",
       "      <td>154823</td>\n",
       "      <td>54</td>\n",
       "      <td>214</td>\n",
       "      <td>124</td>\n",
       "      <td>0.077694</td>\n",
       "      <td>0.435484</td>\n",
       "      <td>0.033835</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245654</th>\n",
       "      <td>103602</td>\n",
       "      <td>32</td>\n",
       "      <td>52343</td>\n",
       "      <td>32</td>\n",
       "      <td>489</td>\n",
       "      <td>32</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245655</th>\n",
       "      <td>103603</td>\n",
       "      <td>24</td>\n",
       "      <td>73469</td>\n",
       "      <td>24</td>\n",
       "      <td>545</td>\n",
       "      <td>24</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245656</th>\n",
       "      <td>103606</td>\n",
       "      <td>4</td>\n",
       "      <td>124997</td>\n",
       "      <td>2</td>\n",
       "      <td>693</td>\n",
       "      <td>2</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245657</th>\n",
       "      <td>103606</td>\n",
       "      <td>4</td>\n",
       "      <td>146704</td>\n",
       "      <td>2</td>\n",
       "      <td>800</td>\n",
       "      <td>2</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245658</th>\n",
       "      <td>103607</td>\n",
       "      <td>360</td>\n",
       "      <td>5505</td>\n",
       "      <td>13</td>\n",
       "      <td>214</td>\n",
       "      <td>249</td>\n",
       "      <td>0.691667</td>\n",
       "      <td>0.052209</td>\n",
       "      <td>0.036111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245659</th>\n",
       "      <td>103607</td>\n",
       "      <td>360</td>\n",
       "      <td>18412</td>\n",
       "      <td>44</td>\n",
       "      <td>214</td>\n",
       "      <td>249</td>\n",
       "      <td>0.691667</td>\n",
       "      <td>0.176707</td>\n",
       "      <td>0.122222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245660</th>\n",
       "      <td>103607</td>\n",
       "      <td>360</td>\n",
       "      <td>32465</td>\n",
       "      <td>52</td>\n",
       "      <td>489</td>\n",
       "      <td>97</td>\n",
       "      <td>0.269444</td>\n",
       "      <td>0.536082</td>\n",
       "      <td>0.144444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245661</th>\n",
       "      <td>103607</td>\n",
       "      <td>360</td>\n",
       "      <td>61531</td>\n",
       "      <td>20</td>\n",
       "      <td>214</td>\n",
       "      <td>249</td>\n",
       "      <td>0.691667</td>\n",
       "      <td>0.080321</td>\n",
       "      <td>0.055556</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245662</th>\n",
       "      <td>103607</td>\n",
       "      <td>360</td>\n",
       "      <td>63006</td>\n",
       "      <td>158</td>\n",
       "      <td>214</td>\n",
       "      <td>249</td>\n",
       "      <td>0.691667</td>\n",
       "      <td>0.634538</td>\n",
       "      <td>0.438889</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245663</th>\n",
       "      <td>103607</td>\n",
       "      <td>360</td>\n",
       "      <td>81708</td>\n",
       "      <td>45</td>\n",
       "      <td>489</td>\n",
       "      <td>97</td>\n",
       "      <td>0.269444</td>\n",
       "      <td>0.463918</td>\n",
       "      <td>0.125000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245664</th>\n",
       "      <td>103607</td>\n",
       "      <td>360</td>\n",
       "      <td>112332</td>\n",
       "      <td>7</td>\n",
       "      <td>214</td>\n",
       "      <td>249</td>\n",
       "      <td>0.691667</td>\n",
       "      <td>0.028112</td>\n",
       "      <td>0.019444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245665</th>\n",
       "      <td>103607</td>\n",
       "      <td>360</td>\n",
       "      <td>153698</td>\n",
       "      <td>14</td>\n",
       "      <td>306</td>\n",
       "      <td>14</td>\n",
       "      <td>0.038889</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.038889</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245666</th>\n",
       "      <td>103607</td>\n",
       "      <td>360</td>\n",
       "      <td>169085</td>\n",
       "      <td>7</td>\n",
       "      <td>214</td>\n",
       "      <td>249</td>\n",
       "      <td>0.691667</td>\n",
       "      <td>0.028112</td>\n",
       "      <td>0.019444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245667</th>\n",
       "      <td>103611</td>\n",
       "      <td>22</td>\n",
       "      <td>2677</td>\n",
       "      <td>1</td>\n",
       "      <td>427</td>\n",
       "      <td>1</td>\n",
       "      <td>0.045455</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.045455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245668</th>\n",
       "      <td>103611</td>\n",
       "      <td>22</td>\n",
       "      <td>42721</td>\n",
       "      <td>7</td>\n",
       "      <td>30</td>\n",
       "      <td>14</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.318182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245669</th>\n",
       "      <td>103611</td>\n",
       "      <td>22</td>\n",
       "      <td>65520</td>\n",
       "      <td>7</td>\n",
       "      <td>30</td>\n",
       "      <td>14</td>\n",
       "      <td>0.636364</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.318182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245670</th>\n",
       "      <td>103611</td>\n",
       "      <td>22</td>\n",
       "      <td>166707</td>\n",
       "      <td>7</td>\n",
       "      <td>489</td>\n",
       "      <td>7</td>\n",
       "      <td>0.318182</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.318182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245671</th>\n",
       "      <td>103612</td>\n",
       "      <td>190</td>\n",
       "      <td>63006</td>\n",
       "      <td>42</td>\n",
       "      <td>214</td>\n",
       "      <td>164</td>\n",
       "      <td>0.863158</td>\n",
       "      <td>0.256098</td>\n",
       "      <td>0.221053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245672</th>\n",
       "      <td>103612</td>\n",
       "      <td>190</td>\n",
       "      <td>105559</td>\n",
       "      <td>20</td>\n",
       "      <td>214</td>\n",
       "      <td>164</td>\n",
       "      <td>0.863158</td>\n",
       "      <td>0.121951</td>\n",
       "      <td>0.105263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245673</th>\n",
       "      <td>103612</td>\n",
       "      <td>190</td>\n",
       "      <td>116746</td>\n",
       "      <td>76</td>\n",
       "      <td>214</td>\n",
       "      <td>164</td>\n",
       "      <td>0.863158</td>\n",
       "      <td>0.463415</td>\n",
       "      <td>0.400000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245674</th>\n",
       "      <td>103612</td>\n",
       "      <td>190</td>\n",
       "      <td>128167</td>\n",
       "      <td>26</td>\n",
       "      <td>214</td>\n",
       "      <td>164</td>\n",
       "      <td>0.863158</td>\n",
       "      <td>0.158537</td>\n",
       "      <td>0.136842</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245675</th>\n",
       "      <td>103612</td>\n",
       "      <td>190</td>\n",
       "      <td>154636</td>\n",
       "      <td>26</td>\n",
       "      <td>545</td>\n",
       "      <td>26</td>\n",
       "      <td>0.136842</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.136842</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245676</th>\n",
       "      <td>103613</td>\n",
       "      <td>233</td>\n",
       "      <td>18412</td>\n",
       "      <td>55</td>\n",
       "      <td>214</td>\n",
       "      <td>55</td>\n",
       "      <td>0.236052</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.236052</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245677</th>\n",
       "      <td>103613</td>\n",
       "      <td>233</td>\n",
       "      <td>21147</td>\n",
       "      <td>25</td>\n",
       "      <td>403</td>\n",
       "      <td>109</td>\n",
       "      <td>0.467811</td>\n",
       "      <td>0.229358</td>\n",
       "      <td>0.107296</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245678</th>\n",
       "      <td>103613</td>\n",
       "      <td>233</td>\n",
       "      <td>126535</td>\n",
       "      <td>13</td>\n",
       "      <td>545</td>\n",
       "      <td>56</td>\n",
       "      <td>0.240343</td>\n",
       "      <td>0.232143</td>\n",
       "      <td>0.055794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245679</th>\n",
       "      <td>103613</td>\n",
       "      <td>233</td>\n",
       "      <td>134084</td>\n",
       "      <td>13</td>\n",
       "      <td>693</td>\n",
       "      <td>13</td>\n",
       "      <td>0.055794</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.055794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245680</th>\n",
       "      <td>103613</td>\n",
       "      <td>233</td>\n",
       "      <td>138778</td>\n",
       "      <td>39</td>\n",
       "      <td>403</td>\n",
       "      <td>109</td>\n",
       "      <td>0.467811</td>\n",
       "      <td>0.357798</td>\n",
       "      <td>0.167382</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245681</th>\n",
       "      <td>103613</td>\n",
       "      <td>233</td>\n",
       "      <td>147833</td>\n",
       "      <td>32</td>\n",
       "      <td>403</td>\n",
       "      <td>109</td>\n",
       "      <td>0.467811</td>\n",
       "      <td>0.293578</td>\n",
       "      <td>0.137339</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245682</th>\n",
       "      <td>103613</td>\n",
       "      <td>233</td>\n",
       "      <td>151327</td>\n",
       "      <td>13</td>\n",
       "      <td>403</td>\n",
       "      <td>109</td>\n",
       "      <td>0.467811</td>\n",
       "      <td>0.119266</td>\n",
       "      <td>0.055794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245683</th>\n",
       "      <td>103613</td>\n",
       "      <td>233</td>\n",
       "      <td>154636</td>\n",
       "      <td>43</td>\n",
       "      <td>545</td>\n",
       "      <td>56</td>\n",
       "      <td>0.240343</td>\n",
       "      <td>0.767857</td>\n",
       "      <td>0.184549</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>245684 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        user_id  user_cnt  item_id  item_cnt  item_category  category_cnt  \\\n",
       "0             1        76   124997        76            693            76   \n",
       "1             8         6    96887         6            214             6   \n",
       "2            11        90    52343        32            489            32   \n",
       "3            11        90    57018        58            545            58   \n",
       "4            15        90    18412        90            214            90   \n",
       "5            17       109      285        13            214            71   \n",
       "6            17       109    21457        13            214            71   \n",
       "7            17       109    58959        38            622            38   \n",
       "8            17       109    60077        13            214            71   \n",
       "9            17       109    65428        13            214            71   \n",
       "10           17       109   107318        13            214            71   \n",
       "11           17       109   151195         6            214            71   \n",
       "12           22      1596    15259       188            545          1438   \n",
       "13           22      1596    24371        36            545          1438   \n",
       "14           22      1596    24423        26            214           124   \n",
       "15           22      1596    47950        56            545          1438   \n",
       "16           22      1596    48154        44            214           124   \n",
       "17           22      1596    48895        34            489            34   \n",
       "18           22      1596    49178       124            545          1438   \n",
       "19           22      1596    56234        38            545          1438   \n",
       "20           22      1596    75836        28            545          1438   \n",
       "21           22      1596    78720        64            545          1438   \n",
       "22           22      1596    89802        38            545          1438   \n",
       "23           22      1596   108824        66            545          1438   \n",
       "24           22      1596   110779        52            545          1438   \n",
       "25           22      1596   116489        26            545          1438   \n",
       "26           22      1596   128988        66            545          1438   \n",
       "27           22      1596   132562        52            545          1438   \n",
       "28           22      1596   146634        26            545          1438   \n",
       "29           22      1596   154823        54            214           124   \n",
       "...         ...       ...      ...       ...            ...           ...   \n",
       "245654   103602        32    52343        32            489            32   \n",
       "245655   103603        24    73469        24            545            24   \n",
       "245656   103606         4   124997         2            693             2   \n",
       "245657   103606         4   146704         2            800             2   \n",
       "245658   103607       360     5505        13            214           249   \n",
       "245659   103607       360    18412        44            214           249   \n",
       "245660   103607       360    32465        52            489            97   \n",
       "245661   103607       360    61531        20            214           249   \n",
       "245662   103607       360    63006       158            214           249   \n",
       "245663   103607       360    81708        45            489            97   \n",
       "245664   103607       360   112332         7            214           249   \n",
       "245665   103607       360   153698        14            306            14   \n",
       "245666   103607       360   169085         7            214           249   \n",
       "245667   103611        22     2677         1            427             1   \n",
       "245668   103611        22    42721         7             30            14   \n",
       "245669   103611        22    65520         7             30            14   \n",
       "245670   103611        22   166707         7            489             7   \n",
       "245671   103612       190    63006        42            214           164   \n",
       "245672   103612       190   105559        20            214           164   \n",
       "245673   103612       190   116746        76            214           164   \n",
       "245674   103612       190   128167        26            214           164   \n",
       "245675   103612       190   154636        26            545            26   \n",
       "245676   103613       233    18412        55            214            55   \n",
       "245677   103613       233    21147        25            403           109   \n",
       "245678   103613       233   126535        13            545            56   \n",
       "245679   103613       233   134084        13            693            13   \n",
       "245680   103613       233   138778        39            403           109   \n",
       "245681   103613       233   147833        32            403           109   \n",
       "245682   103613       233   151327        13            403           109   \n",
       "245683   103613       233   154636        43            545            56   \n",
       "\n",
       "        category_user_pcn  item_category_pcn  item_user_pcn  \n",
       "0                1.000000           1.000000       1.000000  \n",
       "1                1.000000           1.000000       1.000000  \n",
       "2                0.355556           1.000000       0.355556  \n",
       "3                0.644444           1.000000       0.644444  \n",
       "4                1.000000           1.000000       1.000000  \n",
       "5                0.651376           0.183099       0.119266  \n",
       "6                0.651376           0.183099       0.119266  \n",
       "7                0.348624           1.000000       0.348624  \n",
       "8                0.651376           0.183099       0.119266  \n",
       "9                0.651376           0.183099       0.119266  \n",
       "10               0.651376           0.183099       0.119266  \n",
       "11               0.651376           0.084507       0.055046  \n",
       "12               0.901003           0.130737       0.117794  \n",
       "13               0.901003           0.025035       0.022556  \n",
       "14               0.077694           0.209677       0.016291  \n",
       "15               0.901003           0.038943       0.035088  \n",
       "16               0.077694           0.354839       0.027569  \n",
       "17               0.021303           1.000000       0.021303  \n",
       "18               0.901003           0.086231       0.077694  \n",
       "19               0.901003           0.026426       0.023810  \n",
       "20               0.901003           0.019471       0.017544  \n",
       "21               0.901003           0.044506       0.040100  \n",
       "22               0.901003           0.026426       0.023810  \n",
       "23               0.901003           0.045897       0.041353  \n",
       "24               0.901003           0.036161       0.032581  \n",
       "25               0.901003           0.018081       0.016291  \n",
       "26               0.901003           0.045897       0.041353  \n",
       "27               0.901003           0.036161       0.032581  \n",
       "28               0.901003           0.018081       0.016291  \n",
       "29               0.077694           0.435484       0.033835  \n",
       "...                   ...                ...            ...  \n",
       "245654           1.000000           1.000000       1.000000  \n",
       "245655           1.000000           1.000000       1.000000  \n",
       "245656           0.500000           1.000000       0.500000  \n",
       "245657           0.500000           1.000000       0.500000  \n",
       "245658           0.691667           0.052209       0.036111  \n",
       "245659           0.691667           0.176707       0.122222  \n",
       "245660           0.269444           0.536082       0.144444  \n",
       "245661           0.691667           0.080321       0.055556  \n",
       "245662           0.691667           0.634538       0.438889  \n",
       "245663           0.269444           0.463918       0.125000  \n",
       "245664           0.691667           0.028112       0.019444  \n",
       "245665           0.038889           1.000000       0.038889  \n",
       "245666           0.691667           0.028112       0.019444  \n",
       "245667           0.045455           1.000000       0.045455  \n",
       "245668           0.636364           0.500000       0.318182  \n",
       "245669           0.636364           0.500000       0.318182  \n",
       "245670           0.318182           1.000000       0.318182  \n",
       "245671           0.863158           0.256098       0.221053  \n",
       "245672           0.863158           0.121951       0.105263  \n",
       "245673           0.863158           0.463415       0.400000  \n",
       "245674           0.863158           0.158537       0.136842  \n",
       "245675           0.136842           1.000000       0.136842  \n",
       "245676           0.236052           1.000000       0.236052  \n",
       "245677           0.467811           0.229358       0.107296  \n",
       "245678           0.240343           0.232143       0.055794  \n",
       "245679           0.055794           1.000000       0.055794  \n",
       "245680           0.467811           0.357798       0.167382  \n",
       "245681           0.467811           0.293578       0.137339  \n",
       "245682           0.467811           0.119266       0.055794  \n",
       "245683           0.240343           0.767857       0.184549  \n",
       "\n",
       "[245684 rows x 9 columns]"
      ]
     },
     "execution_count": 150,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "usm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>user_id</th>\n",
       "      <th>item_id</th>\n",
       "      <th>behavior_type</th>\n",
       "      <th>brand</th>\n",
       "      <th>date</th>\n",
       "      <th>hour</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>208266</td>\n",
       "      <td>31662</td>\n",
       "      <td>1</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>209390</td>\n",
       "      <td>118799</td>\n",
       "      <td>6</td>\n",
       "      <td>244</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>237311</td>\n",
       "      <td>5825</td>\n",
       "      <td>1</td>\n",
       "      <td>885</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>297325</td>\n",
       "      <td>128747</td>\n",
       "      <td>6</td>\n",
       "      <td>489</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>287896</td>\n",
       "      <td>68767</td>\n",
       "      <td>6</td>\n",
       "      <td>545</td>\n",
       "      <td>2016-03-31</td>\n",
       "      <td>23:59:07</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   user_id  item_id  behavior_type  brand        date      hour\n",
       "0   208266    31662              1    545  2016-03-31  23:59:04\n",
       "1   209390   118799              6    244  2016-03-31  23:59:05\n",
       "2   237311     5825              1    885  2016-03-31  23:59:05\n",
       "3   297325   128747              6    489  2016-03-31  23:59:06\n",
       "4   287896    68767              6    545  2016-03-31  23:59:07"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "u.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "check_day = u[u.date.ge('2016-04-11')&u.date.lt('2016-04-16')&u.behavior_type.eq(4)].user_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "check_day_no = u[u.date.ge('2016-04-11')&u.date.lt('2016-04-16')&u.behavior_type.ne(4)].user_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "before_day = u[u.date.lt('2016-04-11')&u.date.ge('2016-04-03')].user_id.isin(check_day) # 在后五天有购买的用户"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "before_info = u[u.date.lt('2016-04-11')&u.date.ge('2016-04-03')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "act =pd.get_dummies(before_info.behavior_type,prefix='action')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "all = pd.concat([before_info, act], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "buy = all[before_day].groupby(['user_id']).sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "invalid syntax (<ipython-input-43-f66e6bbbf2be>, line 1)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;36m  File \u001b[1;32m\"<ipython-input-43-f66e6bbbf2be>\"\u001b[1;36m, line \u001b[1;32m1\u001b[0m\n\u001b[1;33m    浏览,加入购物车，购物车删除，下单，关注，点击\u001b[0m\n\u001b[1;37m    ^\u001b[0m\n\u001b[1;31mSyntaxError\u001b[0m\u001b[1;31m:\u001b[0m invalid syntax\n"
     ]
    }
   ],
   "source": [
    "浏览,加入购物车，购物车删除，下单，关注，点击"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>item_id</th>\n",
       "      <th>behavior_type</th>\n",
       "      <th>brand</th>\n",
       "      <th>action_1</th>\n",
       "      <th>action_2</th>\n",
       "      <th>action_3</th>\n",
       "      <th>action_4</th>\n",
       "      <th>action_5</th>\n",
       "      <th>action_6</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>user_id</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>206931</th>\n",
       "      <td>1741136</td>\n",
       "      <td>56</td>\n",
       "      <td>6072</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206951</th>\n",
       "      <td>3383382</td>\n",
       "      <td>175</td>\n",
       "      <td>24464</td>\n",
       "      <td>19.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206991</th>\n",
       "      <td>1538122</td>\n",
       "      <td>54</td>\n",
       "      <td>420</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207052</th>\n",
       "      <td>2752552</td>\n",
       "      <td>167</td>\n",
       "      <td>20165</td>\n",
       "      <td>7.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>25.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207056</th>\n",
       "      <td>15981048</td>\n",
       "      <td>618</td>\n",
       "      <td>66382</td>\n",
       "      <td>55.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>91.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207060</th>\n",
       "      <td>1214532</td>\n",
       "      <td>53</td>\n",
       "      <td>8802</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207077</th>\n",
       "      <td>31662</td>\n",
       "      <td>1</td>\n",
       "      <td>545</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207106</th>\n",
       "      <td>5825</td>\n",
       "      <td>1</td>\n",
       "      <td>885</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207143</th>\n",
       "      <td>32207587</td>\n",
       "      <td>969</td>\n",
       "      <td>148598</td>\n",
       "      <td>122.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>137.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207149</th>\n",
       "      <td>327684</td>\n",
       "      <td>4</td>\n",
       "      <td>1304</td>\n",
       "      <td>2.0</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207183</th>\n",
       "      <td>1314016</td>\n",
       "      <td>12</td>\n",
       "      <td>6400</td>\n",
       "      <td>5.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207195</th>\n",
       "      <td>2342322</td>\n",
       "      <td>174</td>\n",
       "      <td>15475</td>\n",
       "      <td>18.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207223</th>\n",
       "      <td>2687941</td>\n",
       "      <td>52</td>\n",
       "      <td>2615</td>\n",
       "      <td>18.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207244</th>\n",
       "      <td>2511824</td>\n",
       "      <td>182</td>\n",
       "      <td>1680</td>\n",
       "      <td>30.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>25.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207248</th>\n",
       "      <td>13366369</td>\n",
       "      <td>416</td>\n",
       "      <td>67478</td>\n",
       "      <td>50.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>57.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207262</th>\n",
       "      <td>2003282</td>\n",
       "      <td>245</td>\n",
       "      <td>34334</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>39.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207280</th>\n",
       "      <td>376920</td>\n",
       "      <td>131</td>\n",
       "      <td>6420</td>\n",
       "      <td>9.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>20.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207281</th>\n",
       "      <td>1275516</td>\n",
       "      <td>68</td>\n",
       "      <td>9667</td>\n",
       "      <td>6.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207291</th>\n",
       "      <td>798252</td>\n",
       "      <td>30</td>\n",
       "      <td>7630</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>207369</th>\n",
       "      <td>4467257</td>\n",
       "      <td>124</td>\n",
       "      <td>19735</td>\n",
       "      <td>12.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>18.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          item_id  behavior_type   brand  action_1  action_2  action_3  \\\n",
       "user_id                                                                  \n",
       "206931    1741136             56    6072       8.0       0.0       0.0   \n",
       "206951    3383382            175   24464      19.0       0.0       0.0   \n",
       "206991    1538122             54     420       6.0       0.0       0.0   \n",
       "207052    2752552            167   20165       7.0       5.0       0.0   \n",
       "207056   15981048            618   66382      55.0       1.0       0.0   \n",
       "207060    1214532             53    8802      11.0       0.0       0.0   \n",
       "207077      31662              1     545       1.0       0.0       0.0   \n",
       "207106       5825              1     885       1.0       0.0       0.0   \n",
       "207143   32207587            969  148598     122.0      10.0       0.0   \n",
       "207149     327684              4    1304       2.0       1.0       0.0   \n",
       "207183    1314016             12    6400       5.0       2.0       1.0   \n",
       "207195    2342322            174   15475      18.0       0.0       0.0   \n",
       "207223    2687941             52    2615      18.0       2.0       0.0   \n",
       "207244    2511824            182    1680      30.0       1.0       0.0   \n",
       "207248   13366369            416   67478      50.0       6.0       4.0   \n",
       "207262    2003282            245   34334       8.0       0.0       1.0   \n",
       "207280     376920            131    6420       9.0       1.0       0.0   \n",
       "207281    1275516             68    9667       6.0       1.0       0.0   \n",
       "207291     798252             30    7630      10.0       1.0       0.0   \n",
       "207369    4467257            124   19735      12.0       2.0       0.0   \n",
       "\n",
       "         action_4  action_5  action_6  \n",
       "user_id                                \n",
       "206931        0.0       0.0       8.0  \n",
       "206951        0.0       0.0      26.0  \n",
       "206991        0.0       0.0       8.0  \n",
       "207052        0.0       0.0      25.0  \n",
       "207056        0.0       3.0      91.0  \n",
       "207060        0.0       0.0       7.0  \n",
       "207077        0.0       0.0       0.0  \n",
       "207106        0.0       0.0       0.0  \n",
       "207143        0.0       1.0     137.0  \n",
       "207149        0.0       0.0       0.0  \n",
       "207183        0.0       0.0       0.0  \n",
       "207195        0.0       0.0      26.0  \n",
       "207223        0.0       0.0       5.0  \n",
       "207244        0.0       0.0      25.0  \n",
       "207248        0.0       0.0      57.0  \n",
       "207262        0.0       0.0      39.0  \n",
       "207280        0.0       0.0      20.0  \n",
       "207281        0.0       0.0      10.0  \n",
       "207291        0.0       0.0       3.0  \n",
       "207369        0.0       0.0      18.0  "
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "buy.iloc[50:70]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "no_buy = all[~before_day].groupby(['user_id']).sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "no_buy.drop(['item_id','behavior_type','brand'], axis = 1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "no_buy.columns = ['浏览','加入购物车','购物车删除','下单','关注','点击']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "??pd.Series.sort_values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "user_id\n",
       "293215    1481.0\n",
       "247006     914.0\n",
       "291207     826.0\n",
       "231288     783.0\n",
       "209147     782.0\n",
       "258626     772.0\n",
       "227874     738.0\n",
       "214315     729.0\n",
       "258487     712.0\n",
       "284730     709.0\n",
       "271506     708.0\n",
       "223340     646.0\n",
       "274217     644.0\n",
       "267848     637.0\n",
       "212356     626.0\n",
       "213006     578.0\n",
       "229039     576.0\n",
       "295193     576.0\n",
       "257176     572.0\n",
       "207069     530.0\n",
       "293301     524.0\n",
       "280079     518.0\n",
       "211094     518.0\n",
       "214658     516.0\n",
       "261135     507.0\n",
       "223181     505.0\n",
       "217174     504.0\n",
       "259211     494.0\n",
       "216146     492.0\n",
       "239431     487.0\n",
       "           ...  \n",
       "271993       0.0\n",
       "212489       0.0\n",
       "212488       0.0\n",
       "245483       0.0\n",
       "271980       0.0\n",
       "212471       0.0\n",
       "291142       0.0\n",
       "271959       0.0\n",
       "291167       0.0\n",
       "291180       0.0\n",
       "231556       0.0\n",
       "271790       0.0\n",
       "245445       0.0\n",
       "245429       0.0\n",
       "212390       0.0\n",
       "212378       0.0\n",
       "271740       0.0\n",
       "212358       0.0\n",
       "212335       0.0\n",
       "271690       0.0\n",
       "231707       0.0\n",
       "257939       0.0\n",
       "231714       0.0\n",
       "231730       0.0\n",
       "212283       0.0\n",
       "245357       0.0\n",
       "271666       0.0\n",
       "258000       0.0\n",
       "231783       0.0\n",
       "246169       0.0\n",
       "Name: 浏览, dtype: float64"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "no_buy['浏览'].sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>浏览</th>\n",
       "      <th>加入购物车</th>\n",
       "      <th>购物车删除</th>\n",
       "      <th>下单</th>\n",
       "      <th>关注</th>\n",
       "      <th>点击</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>user_id</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>293215</th>\n",
       "      <td>1481.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>2362.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             浏览  加入购物车  购物车删除   下单    关注      点击\n",
       "user_id                                         \n",
       "293215   1481.0    6.0    4.0  0.0  11.0  2362.0"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "no_buy[no_buy['浏览'] == 1481]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "item_id          4.964822e+06\n",
       "behavior_type    2.429730e+02\n",
       "brand            2.619873e+04\n",
       "action_1         2.153418e+01\n",
       "action_2         7.536129e-01\n",
       "action_3         3.048430e-01\n",
       "action_4         4.715062e-02\n",
       "action_5         1.232250e-01\n",
       "action_6         3.636873e+01\n",
       "dtype: float64"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "no_buy.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7536129"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "7.536129e-01"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}