user behaviors

{
 "metadata": {
  "name": ""
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "heading",
     "level": 1,
     "metadata": {},
     "source": [
      "\u4f7f\u7528\u8005\u884c\u70ba\u95dc\u9023\u5206\u6790"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%pylab inline --no-import-all\n",
      "\n",
      "from pylab import *\n",
      "import prettyplotlib as ppl\n",
      "from prettyplotlib import plt\n",
      "import numpy as np\n",
      "import MySQLdb\n",
      "import math\n",
      "from datetime import datetime, timedelta\n",
      "\n",
      "db = MySQLdb.connect(host=\"localhost\", user=\"root\", db=\"cookpad_production\")\n",
      "cur = db.cursor()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Populating the interactive namespace from numpy and matplotlib\n"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "cur.execute(\"\\\n",
      "select b.id, b.ts, b.act from \\\n",
      "    ((select user_id id, created_at ts, 1 as act from favorites) \\\n",
      "    union (select follower_id id, created_at ts, 2 as act from follows) \\\n",
      "    union (select user_id id, created_at ts, 3 as act from recipes)) b order by b.ts \\\n",
      "            \")\n",
      "\n",
      "behaviors = cur.fetchall()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# counter = 0\n",
      "user_seq_behavior = {}\n",
      "for behavior in behaviors:\n",
      "    #if counter == 10000:\n",
      "    #    break\n",
      "    #else:\n",
      "    #    counter += 1\n",
      "    # print str(behavior[0]) + u'\\t' + str(behavior[1] + timedelta(hours=8)) + u'\\t' + str(behavior[2])\n",
      "    if behavior[0] in user_seq_behavior:\n",
      "        if user_seq_behavior[behavior[0]][len(user_seq_behavior[behavior[0]])-1] != behavior[2]:\n",
      "            user_seq_behavior[behavior[0]].append(behavior[2]) \n",
      "    else:\n",
      "        user_seq_behavior[behavior[0]] = [behavior[2]]\n",
      "        "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 5
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "result = numpy.zeros(9).reshape((3, 3))\n",
      "total = [0, 0, 0]\n",
      "for uid in user_seq_behavior.keys():\n",
      "    tmp = -1\n",
      "    for i in range(len(user_seq_behavior[uid])):\n",
      "        if tmp != -1:\n",
      "            result[tmp][user_seq_behavior[uid][i]-1] += 1\n",
      "        tmp = user_seq_behavior[uid][i]-1\n",
      "        total[user_seq_behavior[uid][i]-1] += 1\n",
      "\n",
      "print result\n",
      "print total"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[[     0.  38619.   9231.]\n",
        " [ 96944.      0.   5436.]\n",
        " [ 10944.   1243.      0.]]\n",
        "[230535, 168099, 16896]\n"
       ]
      }
     ],
     "prompt_number": 15
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print u'\u6536\u85cf\u5f8c\u8ffd\u8e64\u7684\u6a5f\u7387 ' + str(result[0][1]/total[0])\n",
      "print u'\u6536\u85cf\u5f8c\u64b0\u5beb\u7684\u6a5f\u7387 ' + str(result[0][2]/total[0])\n",
      "print u'\u8ffd\u8e64\u5f8c\u6536\u85cf\u7684\u6a5f\u7387 ' + str(result[1][0]/total[1])\n",
      "print u'\u8ffd\u8e64\u5f8c\u64b0\u5beb\u7684\u6a5f\u7387 ' + str(result[1][2]/total[1])\n",
      "print u'\u64b0\u5beb\u5f8c\u6536\u85cf\u7684\u6a5f\u7387 ' + str(result[2][0]/total[2])\n",
      "print u'\u64b0\u5beb\u5f8c\u8ffd\u8e64\u7684\u6a5f\u7387 ' + str(result[2][1]/total[2])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\u6536\u85cf\u5f8c\u8ffd\u8e64\u7684\u6a5f\u7387 0.167519031817\n",
        "\u6536\u85cf\u5f8c\u64b0\u5beb\u7684\u6a5f\u7387 0.0400416422669\n",
        "\u8ffd\u8e64\u5f8c\u6536\u85cf\u7684\u6a5f\u7387 0.576707773395\n",
        "\u8ffd\u8e64\u5f8c\u64b0\u5beb\u7684\u6a5f\u7387 0.0323380864848\n",
        "\u64b0\u5beb\u5f8c\u6536\u85cf\u7684\u6a5f\u7387 0.647727272727\n",
        "\u64b0\u5beb\u5f8c\u8ffd\u8e64\u7684\u6a5f\u7387 0.0735677083333\n"
       ]
      }
     ],
     "prompt_number": 17
    }
   ],
   "metadata": {}
  }
 ]
}