Art Features Analysis.ipynb

{
 "metadata": {
  "name": ""
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%matplotlib inline\n",
      "import csv\n",
      "import pandas as pd\n",
      "import numpy as np\n",
      "import matplotlib.pyplot as plt\n",
      "from sklearn.cluster import KMeans\n",
      "from sklearn.preprocessing import normalize\n",
      "from img_feature_extraction import img_from_url\n",
      "import scipy.spatial as sp\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "art = pd.DataFrame.from_csv('data/artwork_features_test.csv', index_col= [0,1])\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "art.head()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th></th>\n",
        "      <th>avg_blue</th>\n",
        "      <th>high_blue</th>\n",
        "      <th>med_blue</th>\n",
        "      <th>low_blue</th>\n",
        "      <th>avg_green</th>\n",
        "      <th>high_green</th>\n",
        "      <th>med_green</th>\n",
        "      <th>low_green</th>\n",
        "      <th>avg_red</th>\n",
        "      <th>high_red</th>\n",
        "      <th>med_red</th>\n",
        "      <th>low_red</th>\n",
        "      <th>avg_gray</th>\n",
        "      <th>high_gray</th>\n",
        "      <th>med_gray</th>\n",
        "      <th>low_gray</th>\n",
        "      <th>avg_l</th>\n",
        "      <th>high_l</th>\n",
        "      <th>med_l</th>\n",
        "      <th>low_l</th>\n",
        "      <th></th>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th></th>\n",
        "      <th>Unnamed: 1</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",
        "      <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",
        "      <th></th>\n",
        "      <th></th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>1-681</th>\n",
        "      <th>http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg</th>\n",
        "      <td> 0.769806</td>\n",
        "      <td> 0.688344</td>\n",
        "      <td> 0.193272</td>\n",
        "      <td> 0.118384</td>\n",
        "      <td> 0.764103</td>\n",
        "      <td> 0.678195</td>\n",
        "      <td> 0.197998</td>\n",
        "      <td> 0.123807</td>\n",
        "      <td> 0.766170</td>\n",
        "      <td> 0.682125</td>\n",
        "      <td> 0.196271</td>\n",
        "      <td> 0.121604</td>\n",
        "      <td> 0.766693</td>\n",
        "      <td> 0.682888</td>\n",
        "      <td> 0.195847</td>\n",
        "      <td> 0.121265</td>\n",
        "      <td> 0.766170</td>\n",
        "      <td> 0.682125</td>\n",
        "      <td> 0.196271</td>\n",
        "      <td> 0.121604</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>100taur</th>\n",
        "      <th>http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg</th>\n",
        "      <td> 0.368579</td>\n",
        "      <td> 0.088746</td>\n",
        "      <td> 0.455290</td>\n",
        "      <td> 0.455964</td>\n",
        "      <td> 0.391494</td>\n",
        "      <td> 0.114962</td>\n",
        "      <td> 0.441623</td>\n",
        "      <td> 0.443414</td>\n",
        "      <td> 0.377093</td>\n",
        "      <td> 0.140769</td>\n",
        "      <td> 0.396269</td>\n",
        "      <td> 0.462962</td>\n",
        "      <td> 0.379055</td>\n",
        "      <td> 0.114826</td>\n",
        "      <td> 0.431061</td>\n",
        "      <td> 0.454114</td>\n",
        "      <td> 0.377093</td>\n",
        "      <td> 0.140769</td>\n",
        "      <td> 0.396269</td>\n",
        "      <td> 0.462962</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>10orioart</th>\n",
        "      <th>http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg</th>\n",
        "      <td> 0.249963</td>\n",
        "      <td> 0.000057</td>\n",
        "      <td> 0.557772</td>\n",
        "      <td> 0.442171</td>\n",
        "      <td> 0.331738</td>\n",
        "      <td> 0.000500</td>\n",
        "      <td> 0.712056</td>\n",
        "      <td> 0.287444</td>\n",
        "      <td> 0.372379</td>\n",
        "      <td> 0.002663</td>\n",
        "      <td> 0.729254</td>\n",
        "      <td> 0.268082</td>\n",
        "      <td> 0.318027</td>\n",
        "      <td> 0.001073</td>\n",
        "      <td> 0.666361</td>\n",
        "      <td> 0.332566</td>\n",
        "      <td> 0.372379</td>\n",
        "      <td> 0.002663</td>\n",
        "      <td> 0.729254</td>\n",
        "      <td> 0.268082</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>0rpheus</th>\n",
        "      <th>http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg</th>\n",
        "      <td> 0.333558</td>\n",
        "      <td> 0.008366</td>\n",
        "      <td> 0.653423</td>\n",
        "      <td> 0.338211</td>\n",
        "      <td> 0.403003</td>\n",
        "      <td> 0.024019</td>\n",
        "      <td> 0.709079</td>\n",
        "      <td> 0.266902</td>\n",
        "      <td> 0.467304</td>\n",
        "      <td> 0.157041</td>\n",
        "      <td> 0.594272</td>\n",
        "      <td> 0.248686</td>\n",
        "      <td> 0.401288</td>\n",
        "      <td> 0.063142</td>\n",
        "      <td> 0.652258</td>\n",
        "      <td> 0.284600</td>\n",
        "      <td> 0.467304</td>\n",
        "      <td> 0.157041</td>\n",
        "      <td> 0.594272</td>\n",
        "      <td> 0.248686</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1000scientists</th>\n",
        "      <th>http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg</th>\n",
        "      <td> 0.384384</td>\n",
        "      <td> 0.289747</td>\n",
        "      <td> 0.129810</td>\n",
        "      <td> 0.580443</td>\n",
        "      <td> 0.368052</td>\n",
        "      <td> 0.230152</td>\n",
        "      <td> 0.198803</td>\n",
        "      <td> 0.571046</td>\n",
        "      <td> 0.358938</td>\n",
        "      <td> 0.229769</td>\n",
        "      <td> 0.211240</td>\n",
        "      <td> 0.558991</td>\n",
        "      <td> 0.370458</td>\n",
        "      <td> 0.249889</td>\n",
        "      <td> 0.179951</td>\n",
        "      <td> 0.570160</td>\n",
        "      <td> 0.358938</td>\n",
        "      <td> 0.229769</td>\n",
        "      <td> 0.211240</td>\n",
        "      <td> 0.558991</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "<p>5 rows \u00d7 32 columns</p>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 3,
       "text": [
        "                                                                                                                avg_blue  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.769806   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.368579   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.249963   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.333558   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.384384   \n",
        "\n",
        "                                                                                                                high_blue  \\\n",
        "               Unnamed: 1                                                                                                   \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg   0.688344   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg   0.088746   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg   0.000057   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                    0.008366   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg   0.289747   \n",
        "\n",
        "                                                                                                                med_blue  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.193272   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.455290   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.557772   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.653423   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.129810   \n",
        "\n",
        "                                                                                                                low_blue  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.118384   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.455964   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.442171   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.338211   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.580443   \n",
        "\n",
        "                                                                                                                avg_green  \\\n",
        "               Unnamed: 1                                                                                                   \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg   0.764103   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg   0.391494   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg   0.331738   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                    0.403003   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg   0.368052   \n",
        "\n",
        "                                                                                                                high_green  \\\n",
        "               Unnamed: 1                                                                                                    \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg    0.678195   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg    0.114962   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg    0.000500   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                     0.024019   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg    0.230152   \n",
        "\n",
        "                                                                                                                med_green  \\\n",
        "               Unnamed: 1                                                                                                   \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg   0.197998   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg   0.441623   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg   0.712056   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                    0.709079   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg   0.198803   \n",
        "\n",
        "                                                                                                                low_green  \\\n",
        "               Unnamed: 1                                                                                                   \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg   0.123807   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg   0.443414   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg   0.287444   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                    0.266902   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg   0.571046   \n",
        "\n",
        "                                                                                                                 avg_red  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.766170   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.377093   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.372379   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.467304   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.358938   \n",
        "\n",
        "                                                                                                                high_red  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.682125   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.140769   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.002663   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.157041   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.229769   \n",
        "\n",
        "                                                                                                                 med_red  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.196271   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.396269   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.729254   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.594272   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.211240   \n",
        "\n",
        "                                                                                                                 low_red  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.121604   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.462962   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.268082   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.248686   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.558991   \n",
        "\n",
        "                                                                                                                avg_gray  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.766693   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.379055   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.318027   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.401288   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.370458   \n",
        "\n",
        "                                                                                                                high_gray  \\\n",
        "               Unnamed: 1                                                                                                   \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg   0.682888   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg   0.114826   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg   0.001073   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                    0.063142   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg   0.249889   \n",
        "\n",
        "                                                                                                                med_gray  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.195847   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.431061   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.666361   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.652258   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.179951   \n",
        "\n",
        "                                                                                                                low_gray  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.121265   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.454114   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.332566   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.284600   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.570160   \n",
        "\n",
        "                                                                                                                   avg_l  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.766170   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.377093   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.372379   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.467304   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.358938   \n",
        "\n",
        "                                                                                                                  high_l  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.682125   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.140769   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.002663   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.157041   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.229769   \n",
        "\n",
        "                                                                                                                   med_l  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.196271   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.396269   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.729254   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.594272   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.211240   \n",
        "\n",
        "                                                                                                                   low_l  \\\n",
        "               Unnamed: 1                                                                                                  \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg  0.121604   \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg  0.462962   \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg  0.268082   \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                   0.248686   \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg  0.558991   \n",
        "\n",
        "                                                                                                                    \n",
        "               Unnamed: 1                                                                                           \n",
        "1-681          http://40.media.tumblr.com/4ff949649b2b5889c0c741ffb105c8c0/tumblr_ne9rm4u9Ib1rma1hqo1_1280.jpg ...  \n",
        "100taur        http://41.media.tumblr.com/fb070581e64457555a51ef9f85d15347/tumblr_npj8jyF0vj1rgdqsno1_1280.jpg ...  \n",
        "10orioart      http://40.media.tumblr.com/9698e53e715305fcc9794ed423663ad3/tumblr_noxgg2hafF1t1038go1_1280.jpg ...  \n",
        "0rpheus        http://40.media.tumblr.com/tumblr_m9fzjww9oo1qf0eyqo1_1280.jpg                                  ...  \n",
        "1000scientists http://41.media.tumblr.com/ae09430c50effc8d2240ee3b9ec0132c/tumblr_mpnmv1tQNt1qa6hruo1_1280.jpg ...  \n",
        "\n",
        "[5 rows x 32 columns]"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "art.shape"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 7,
       "text": [
        "(174, 32)"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "art_normal = normalize(art, norm='l2', axis=1, copy=True)\n",
      "art_normal"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 29,
       "text": [
        "array([[  2.82269566e-04,   2.52399365e-04,   7.08684043e-05, ...,\n",
        "          4.26444616e-01,   3.89776979e-01,   4.01170852e-04],\n",
        "       [  1.51291794e-04,   3.64280009e-05,   1.86884489e-04, ...,\n",
        "          4.47416343e-01,   3.67784443e-01,   4.99348597e-04],\n",
        "       [  5.82878456e-04,   1.33596149e-07,   1.30064555e-03, ...,\n",
        "          5.50318971e-01,   2.47177165e-01,   5.19168841e-03],\n",
        "       ..., \n",
        "       [  3.43523501e-04,   3.09434287e-04,   1.15817025e-04, ...,\n",
        "          4.10632856e-01,   4.05858055e-01,   6.90139253e-04],\n",
        "       [  1.61748312e-03,   1.57184218e-03,   2.36301504e-04, ...,\n",
        "          4.52464675e-01,   3.62355184e-01,   2.39399299e-03],\n",
        "       [  2.48323836e-04,   2.02999192e-04,   1.29903652e-04, ...,\n",
        "          4.66654341e-01,   3.46909076e-01,   7.89601254e-04]])"
       ]
      }
     ],
     "prompt_number": 29
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# picking best value of k\n",
      "inertia = []\n",
      "for k in range(1,21):\n",
      "    k_test = KMeans(n_clusters=k)\n",
      "    k_test.fit(art)\n",
      "    inertia.append(k_test.inertia_)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 8
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "x = range(1,21)\n",
      "plt.plot(x, inertia)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 19,
       "text": [
        "[<matplotlib.lines.Line2D at 0x7f42077c5b10>]"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAEGCAYAAACaSwWnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nLiJiIf8Hw+DCVGdTnd4AAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7f424d117b10>"
       ]
      }
     ],
     "prompt_number": 19
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# run r means model\n",
      "k_means = KMeans(n_clusters=4)\n",
      "k_means.fit(art_normal)\n",
      "# dir(k_means)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 36,
       "text": [
        "KMeans(copy_x=True, init='k-means++', max_iter=300, n_clusters=4, n_init=10,\n",
        "    n_jobs=1, precompute_distances='auto', random_state=None, tol=0.0001,\n",
        "    verbose=0)"
       ]
      }
     ],
     "prompt_number": 36
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# get the cluster centers for each of the features\n",
      "k_means.cluster_centers_\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 37,
       "text": [
        "array([[  9.87713014e-04,   4.25105535e-04,   1.64221649e-03,\n",
        "          7.11143094e-04,   1.94104322e-03,   2.05455587e-03,\n",
        "          1.60671819e-04,   5.63237428e-04,   1.97733975e-03,\n",
        "          2.07664580e-03,   1.39812433e-04,   5.62006891e-04,\n",
        "          1.63536533e-03,   1.51876907e-03,   6.47566914e-04,\n",
        "          6.12129138e-04,   1.97733975e-03,   2.07664580e-03,\n",
        "          1.39812433e-04,   5.62006891e-04,   1.94104322e-03,\n",
        "          2.05455587e-03,   1.60671819e-04,   5.63237428e-04,\n",
        "          9.87713014e-04,   4.25105535e-04,   1.64221649e-03,\n",
        "          7.11143094e-04,   7.50126062e-01,   1.05587649e-01,\n",
        "          6.44538412e-01,   6.19788175e-04],\n",
        "       [  7.42490655e-04,   5.76530613e-04,   3.62680763e-04,\n",
        "          2.94425610e-04,   7.95311049e-04,   6.29994233e-04,\n",
        "          3.72661868e-04,   2.30980884e-04,   8.52098256e-04,\n",
        "          7.37978516e-04,   2.87121654e-04,   2.08536816e-04,\n",
        "          7.96633320e-04,   6.48167787e-04,   3.40821428e-04,\n",
        "          2.44647770e-04,   8.52098256e-04,   7.37978516e-04,\n",
        "          2.87121654e-04,   2.08536816e-04,   7.95311049e-04,\n",
        "          6.29994233e-04,   3.72661868e-04,   2.30980884e-04,\n",
        "          7.42490655e-04,   5.76530613e-04,   3.62680763e-04,\n",
        "          2.94425610e-04,   8.14763512e-01,   4.37709844e-01,\n",
        "          3.77053669e-01,   1.45172678e-03],\n",
        "       [  1.11456149e-03,   7.93357923e-04,   7.35737889e-04,\n",
        "          4.87230972e-04,   1.18846344e-03,   8.26101494e-04,\n",
        "          7.93128273e-04,   3.97097017e-04,   1.35999074e-03,\n",
        "          1.22367093e-03,   5.23158352e-04,   2.69497505e-04,\n",
        "          1.22100522e-03,   9.47710115e-04,   6.84008171e-04,\n",
        "          3.84608498e-04,   1.35999074e-03,   1.22367093e-03,\n",
        "          5.23158352e-04,   2.69497505e-04,   1.18846344e-03,\n",
        "          8.26101494e-04,   7.93128273e-04,   3.97097017e-04,\n",
        "          1.11456149e-03,   7.93357923e-04,   7.35737889e-04,\n",
        "          4.87230972e-04,   7.97278287e-01,   5.44401589e-01,\n",
        "          2.52876698e-01,   4.77974614e-03],\n",
        "       [  8.28342455e-04,   6.94673624e-04,   2.80191484e-04,\n",
        "          1.90671415e-04,   8.34544938e-04,   7.17031445e-04,\n",
        "          2.53837299e-04,   1.94667780e-04,   8.48913361e-04,\n",
        "          7.76603861e-04,   1.95795113e-04,   1.93137550e-04,\n",
        "          8.37266918e-04,   7.29436310e-04,   2.43274632e-04,\n",
        "          1.92825582e-04,   8.48913361e-04,   7.76603861e-04,\n",
        "          1.95795113e-04,   1.93137550e-04,   8.34544938e-04,\n",
        "          7.17031445e-04,   2.53837299e-04,   1.94667780e-04,\n",
        "          8.28342455e-04,   6.94673624e-04,   2.80191484e-04,\n",
        "          1.90671415e-04,   8.08913299e-01,   3.16270264e-01,\n",
        "          4.92643034e-01,   8.12258250e-04]])"
       ]
      }
     ],
     "prompt_number": 37
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "col = list(art.columns.values)\n",
      "print len(col)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "32\n"
       ]
      }
     ],
     "prompt_number": 38
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# plot the cluster centers \n",
      "\n",
      "x2 = range(0, len(col))\n",
      "\n",
      "y1 = k_means.cluster_centers_[0]\n",
      "y2 = k_means.cluster_centers_[1]\n",
      "y3 = k_means.cluster_centers_[2]\n",
      "y4 = k_means.cluster_centers_[3]\n",
      "# y5 = k_means.cluster_centers_[4]\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 39
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "\n",
      "plt.plot(x2,y1)\n",
      "plt.plot(x2, y2, color = \"red\")\n",
      "plt.plot(x2, y3, color = \"green\")\n",
      "plt.plot(x2, y4, color = \"orange\")\n",
      "plt.xticks( x2, col )\n",
      "plt.xticks(rotation = 70)\n",
      "# plt.plot(x, y5, color = \"purple\")\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 40,
       "text": [
        "(array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,\n",
        "       17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31]),\n",
        " <a list of 32 Text xticklabel objects>)"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAW8AAAEtCAYAAADHrOFzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7f4207672690>"
       ]
      }
     ],
     "prompt_number": 40
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "k_means.labels_"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 48,
       "text": [
        "array([2, 1, 1, ..., 3, 1, 0], dtype=int32)"
       ]
      }
     ],
     "prompt_number": 48
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# get list indexs with of with level 0\n",
      "labels = k_means.labels_\n",
      "ii = np.where(labels == 0)[0]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 49
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# load url\n",
      "url = \"http://40.media.tumblr.com/bae2ac17797e7eb78038834afa7754e9/tumblr_nhk0vpni6P1teo717o1_1280.jpg\"\n",
      "img = img_from_url(url)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Now extracting features from artwork\n"
       ]
      }
     ],
     "prompt_number": 50
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print img"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[0.63277024089506984, 0.45553608425414366, 0.48365763121546962, 0.060806284530386744, 0.75648071038525055, 0.79441384668508286, 0.20058356353591161, 0.005002589779005525, 0.63882420647042037, 0.22375517955801105, 0.77230058701657456, 0.0039442334254143644]\n"
       ]
      }
     ],
     "prompt_number": 51
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# make predction for img\n",
      "predict = k_means.predict(img)\n",
      "print predict # which cluster does it belong to?\n",
      "cluster = predict[0]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[3]\n"
       ]
      }
     ],
     "prompt_number": 52
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "labels = k_means.labels_\n",
      "# subset dataframe to only include rows in the cluster\n",
      "indexes = np.where(labels == cluster)[0]\n",
      "art_subset = art.iloc[indexes]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 53
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "cluster_center = k_means.cluster_centers_[cluster]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 54
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      " p = art_subset.iloc[0]\n",
      "p "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 55,
       "text": [
        "avg_blue      0.529872\n",
        "high_blue     0.482437\n",
        "med_blue      0.184526\n",
        "low_blue      0.333037\n",
        "avg_green     0.559321\n",
        "high_green    0.617317\n",
        "med_green     0.052295\n",
        "low_green     0.330388\n",
        "avg_red       0.592501\n",
        "high_red      0.637510\n",
        "med_red       0.039990\n",
        "low_red       0.322500\n",
        "Name: (abominationdiary, http://40.media.tumblr.com/ac814254e1e5cf6df4a5d4c0517daf57/tumblr_nmgzhnX8Xz1tssc86o1_1280.jpg), dtype: float64"
       ]
      }
     ],
     "prompt_number": 55
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "nrows = int(art_subset.shape[0])\n",
      "v = img\n",
      "\n",
      "# i = 1\n",
      "# row = list(art_subset.iloc[i])\n",
      "# print row\n",
      "# print img\n",
      "# sp.distance.euclidean(row,v)\n",
      "\n",
      "diff = []\n",
      "\n",
      "for i in range(0,nrows):\n",
      "    row = list(art_subset.iloc[i])\n",
      "    diff.append(sp.distance.euclidean(row,v))\n",
      "art_subset[\"diff\"] = diff\n",
      "# # calculate the euclidean distance from the img vector\n",
      "# def calc_diff(u):\n",
      "#     u = list(u)\n",
      "#     return sp.distance.euclidean(u,v)\n",
      "# art_subset[\"diff\"] =  art_subset.apply(func = calc_diff, axis =1, reduce = True)\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 56
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# sort and fint 5 closest\n",
      "df = art_subset.sort('diff').iloc[0:5]\n",
      "df.reset_index(inplace = True)\n",
      "artists = list(df['X'])\n",
      "art_urls = list(df['X.1'])\n",
      "# get urls they refer to"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 57
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "art_urls"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 58,
       "text": [
        "['http://40.media.tumblr.com/bae2ac17797e7eb78038834afa7754e9/tumblr_nhk0vpni6P1teo717o1_1280.jpg',\n",
        " 'http://41.media.tumblr.com/1f604ffd3567423642e900b2b5e743d9/tumblr_nl1s57xhdl1rt4xnho2_1280.jpg',\n",
        " 'http://41.media.tumblr.com/a99a3ba4d22362a3c0f0cc3f4d7b6c7d/tumblr_njnw0plDrx1qzq15oo1_1280.jpg',\n",
        " 'http://40.media.tumblr.com/tumblr_m9bk1fYs6q1qeju05o1_1280.jpg',\n",
        " 'http://41.media.tumblr.com/fbe1b863983825667f014c98557baa5b/tumblr_nnu2rd8CXE1qkxrtro1_1280.jpg']"
       ]
      }
     ],
     "prompt_number": 58
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}