var_imp.py.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import os\n",
    "import pandas as pd\n",
    "import pandas_td as td\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "con = td.connect(apikey=os.environ['TD_API_KEY'],endpoint='https://api.treasuredata.com/')\n",
    "hive = con.query_engine(database='rossmann', type='hive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div style=\"border-style: dashed; border-width: 1px;\">\n",
       "<div style=\"color: #888;\"># issued at 2016-03-15T06:33:02Z</div>URL: <a href=\"https://console.treasuredata.com/jobs/56302992\" target=\"_blank\">https://console.treasuredata.com/jobs/56302992</a><br>\n",
       "Result size: 179 bytes<br>\n",
       "Download: 179 / 179 bytes (100.00%)<br>\n",
       "<div style=\"color: #888;\"># downloaded at 2016-03-15T06:33:40Z</div></div>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "var_imp=td.read_td('''\n",
    "WITH var_imp AS(\n",
    "  SELECT \n",
    "    array_sum(var_importance) AS var_importance\n",
    "  FROM\n",
    "    model\n",
    ")\n",
    "SELECT \n",
    "  var_importance[0] AS stateholiday,\n",
    "  var_importance[1] AS store,\n",
    "  var_importance[2] AS promo,\n",
    "  var_importance[3] AS dayofweek,\n",
    "  var_importance[4] AS schoolholiday,\n",
    "  var_importance[5] AS promo2sinceweek,\n",
    "  var_importance[6] AS competitionopensinceyear,\n",
    "  var_importance[7] AS assortment,\n",
    "  var_importance[8] AS promo2sinceyear,\n",
    "  var_importance[9] AS promointerval,\n",
    "  var_importance[10] AS promo2,\n",
    "  var_importance[11] AS storetype,\n",
    "  var_importance[12] AS competitionopensincemonth,\n",
    "  var_importance[13] AS YEAR,\n",
    "  var_importance[14] AS MONTH,\n",
    "  var_importance[15] AS DAY,\n",
    "  var_importance[16] AS competitiondistance\n",
    "FROM\n",
    "  var_imp\n",
    "''', hive)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAEICAYAAABGRG3WAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x109a77cd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "X=range(var_imp.shape[1])\n",
    "plt.barh(X,var_imp.values[0],align='center')\n",
    "plt.yticks(X,var_imp.columns)\n",
    "plt.savefig('var_imp.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}