Crustering.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# data analysis and wrangling\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from scipy.stats import lognorm\n",
    "import random as rnd\n",
    "import json\n",
    "import os.path\n",
    "import math\n",
    "from sklearn.metrics import euclidean_distances\n",
    "import face_recognition\n",
    "\n",
    "\n",
    "# visualization\n",
    "import scipy.stats as stats\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.mlab as mlab\n",
    "from PIL import Image\n",
    "# machine learning\n",
    "from sklearn.metrics import silhouette_samples, silhouette_score\n",
    "from sklearn import feature_extraction\n",
    "from sklearn.cluster import KMeans\n",
    "\n",
    "%matplotlib inline\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "VIDEO = 'video_1'\n",
    "PATH  = 'data/{}/data'.format(VIDEO)\n",
    "\n",
    "def imgshow(file):\n",
    "    image = Image.open('data/{}/face/{}'.format(VIDEO, file+\".jpg\"))\n",
    "    image.show()    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Frame_67_Face_74.json\n",
      "Frame_18_Face_18.json\n",
      "Frame_31_Face_27.json\n",
      "Frame_1_Face_1.json\n",
      "Frame_69_Face_75.json\n",
      "Frame_75_Face_81.json\n",
      "Frame_45_Face_48.json\n",
      "Frame_59_Face_64.json\n",
      "Frame_41_Face_40.json\n",
      "Frame_72_Face_78.json\n",
      "Frame_15_Face_15.json\n",
      "Frame_21_Face_22.json\n",
      "Frame_19_Face_19.json\n",
      "Frame_62_Face_67.json\n",
      "Frame_33_Face_29.json\n",
      "Frame_43_Face_45.json\n",
      "Frame_33_Face_28.json\n",
      "Frame_37_Face_36.json\n",
      "Frame_46_Face_50.json\n",
      "Frame_5_Face_5.json\n",
      "Frame_3_Face_3.json\n",
      "Frame_53_Face_59.json\n",
      "Frame_83_Face_86.json\n",
      "Frame_44_Face_47.json\n",
      "Frame_43_Face_44.json\n",
      "Frame_47_Face_53.json\n",
      "Frame_24_Face_25.json\n",
      "Frame_77_Face_83.json\n",
      "Frame_12_Face_12.json\n",
      "Frame_78_Face_84.json\n",
      "Frame_38_Face_37.json\n",
      "Frame_6_Face_6.json\n",
      "Frame_45_Face_49.json\n",
      "Frame_73_Face_79.json\n",
      "Frame_36_Face_33.json\n",
      "Frame_48_Face_54.json\n",
      "Frame_56_Face_61.json\n",
      "Frame_50_Face_58.json\n",
      "Frame_36_Face_34.json\n",
      "Frame_57_Face_62.json\n",
      "Frame_41_Face_41.json\n",
      "Frame_66_Face_71.json\n",
      "Frame_4_Face_4.json\n",
      "Frame_22_Face_24.json\n",
      "Frame_16_Face_16.json\n",
      "Frame_55_Face_60.json\n",
      "Frame_64_Face_69.json\n",
      "Frame_49_Face_57.json\n",
      "Frame_17_Face_17.json\n",
      "Frame_8_Face_8.json\n",
      "Frame_20_Face_21.json\n",
      "Frame_9_Face_9.json\n",
      "Frame_66_Face_72.json\n",
      "Frame_2_Face_2.json\n",
      "Frame_7_Face_7.json\n",
      "Frame_10_Face_10.json\n",
      "Frame_65_Face_70.json\n",
      "Frame_58_Face_63.json\n",
      "Frame_60_Face_65.json\n",
      "Frame_11_Face_11.json\n",
      "Frame_35_Face_32.json\n",
      "Frame_13_Face_13.json\n",
      "Frame_39_Face_39.json\n",
      "Frame_48_Face_55.json\n",
      "Frame_30_Face_26.json\n",
      "Frame_82_Face_85.json\n",
      "Frame_14_Face_14.json\n",
      "Frame_38_Face_38.json\n",
      "Frame_35_Face_31.json\n",
      "Frame_44_Face_46.json\n",
      "Frame_19_Face_20.json\n",
      "Frame_42_Face_43.json\n",
      "Frame_70_Face_76.json\n",
      "Frame_60_Face_66.json\n",
      "Frame_63_Face_68.json\n",
      "Frame_76_Face_82.json\n",
      "Frame_47_Face_52.json\n",
      "Frame_34_Face_30.json\n",
      "Frame_46_Face_51.json\n",
      "Frame_21_Face_23.json\n",
      "Frame_37_Face_35.json\n",
      "Frame_71_Face_77.json\n",
      "Frame_67_Face_73.json\n",
      "Frame_49_Face_56.json\n",
      "Frame_42_Face_42.json\n",
      "Frame_74_Face_80.json\n"
     ]
    }
   ],
   "source": [
    "json_lst = []\n",
    "# Load Files in\n",
    "files = [f for f in os.listdir(PATH)]\n",
    "for f in files:\n",
    "    #print(f)\n",
    "    file_name = PATH+\"/\"+f\n",
    "    data = open(file_name, 'r').read()\n",
    "    data = json.loads(data)\n",
    "    json_lst.append(data)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>age</th>\n",
       "      <th>face_encode</th>\n",
       "      <th>filename</th>\n",
       "      <th>frame</th>\n",
       "      <th>gender</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>28</td>\n",
       "      <td>[-0.06104235351085663, 0.087809719145298, 0.00...</td>\n",
       "      <td>Frame_67_Face_74</td>\n",
       "      <td>67</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>31</td>\n",
       "      <td>[-0.047167032957077026, 0.07096441090106964, 0...</td>\n",
       "      <td>Frame_18_Face_18</td>\n",
       "      <td>18</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>26</td>\n",
       "      <td>[-0.03251942992210388, 0.12871268391609192, 0....</td>\n",
       "      <td>Frame_31_Face_27</td>\n",
       "      <td>31</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>30</td>\n",
       "      <td>[-0.08731923997402191, 0.059576183557510376, 0...</td>\n",
       "      <td>Frame_1_Face_1</td>\n",
       "      <td>1</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>31</td>\n",
       "      <td>[-0.13991206884384155, 0.0679367259144783, 0.0...</td>\n",
       "      <td>Frame_69_Face_75</td>\n",
       "      <td>69</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>27</td>\n",
       "      <td>[-0.020048364996910095, 0.08193252980709076, 0...</td>\n",
       "      <td>Frame_75_Face_81</td>\n",
       "      <td>75</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>32</td>\n",
       "      <td>[-0.058422237634658813, 0.10472062230110168, 0...</td>\n",
       "      <td>Frame_45_Face_48</td>\n",
       "      <td>45</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>37</td>\n",
       "      <td>[-0.08477601408958435, 0.13571122288703918, 0....</td>\n",
       "      <td>Frame_59_Face_64</td>\n",
       "      <td>59</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>25</td>\n",
       "      <td>[-0.06735698878765106, 0.10298667848110199, 0....</td>\n",
       "      <td>Frame_41_Face_40</td>\n",
       "      <td>41</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>31</td>\n",
       "      <td>[-0.08778735995292664, 0.09631244093179703, 0....</td>\n",
       "      <td>Frame_72_Face_78</td>\n",
       "      <td>72</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>29</td>\n",
       "      <td>[-0.0733812004327774, 0.0293489508330822, 0.04...</td>\n",
       "      <td>Frame_15_Face_15</td>\n",
       "      <td>15</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>28</td>\n",
       "      <td>[-0.04642969369888306, 0.08232296258211136, 0....</td>\n",
       "      <td>Frame_21_Face_22</td>\n",
       "      <td>21</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>25</td>\n",
       "      <td>[-0.04680822789669037, 0.07841528952121735, 0....</td>\n",
       "      <td>Frame_19_Face_19</td>\n",
       "      <td>19</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>30</td>\n",
       "      <td>[-0.11061060428619385, 0.12125469744205475, 0....</td>\n",
       "      <td>Frame_62_Face_67</td>\n",
       "      <td>62</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>28</td>\n",
       "      <td>[-0.12499348819255829, 0.1069939061999321, 0.0...</td>\n",
       "      <td>Frame_33_Face_29</td>\n",
       "      <td>33</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>25</td>\n",
       "      <td>[-0.11667513847351074, 0.11083956807851791, 0....</td>\n",
       "      <td>Frame_43_Face_45</td>\n",
       "      <td>43</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>22</td>\n",
       "      <td>[-0.08945988118648529, 0.1424393206834793, 0.0...</td>\n",
       "      <td>Frame_33_Face_28</td>\n",
       "      <td>33</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>23</td>\n",
       "      <td>[-0.12405925989151001, 0.09737150371074677, 0....</td>\n",
       "      <td>Frame_37_Face_36</td>\n",
       "      <td>37</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>32</td>\n",
       "      <td>[-0.05823671817779541, 0.11110067367553711, 0....</td>\n",
       "      <td>Frame_46_Face_50</td>\n",
       "      <td>46</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>28</td>\n",
       "      <td>[-0.06141555309295654, 0.05424799025058746, 0....</td>\n",
       "      <td>Frame_5_Face_5</td>\n",
       "      <td>5</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>30</td>\n",
       "      <td>[-0.07218991219997406, 0.0726621076464653, 0.0...</td>\n",
       "      <td>Frame_3_Face_3</td>\n",
       "      <td>3</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>27</td>\n",
       "      <td>[-0.07461020350456238, 0.08418428897857666, 0....</td>\n",
       "      <td>Frame_53_Face_59</td>\n",
       "      <td>53</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>27</td>\n",
       "      <td>[-0.07565182447433472, 0.07065467536449432, 0....</td>\n",
       "      <td>Frame_83_Face_86</td>\n",
       "      <td>83</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>25</td>\n",
       "      <td>[-0.11129900813102722, 0.15234169363975525, 0....</td>\n",
       "      <td>Frame_44_Face_47</td>\n",
       "      <td>44</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>33</td>\n",
       "      <td>[-0.058597758412361145, 0.0858473852276802, 0....</td>\n",
       "      <td>Frame_43_Face_44</td>\n",
       "      <td>43</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>24</td>\n",
       "      <td>[-0.10831055045127869, 0.14204099774360657, 0....</td>\n",
       "      <td>Frame_47_Face_53</td>\n",
       "      <td>47</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>29</td>\n",
       "      <td>[-0.0750880241394043, 0.06742691993713379, 0.0...</td>\n",
       "      <td>Frame_24_Face_25</td>\n",
       "      <td>24</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>30</td>\n",
       "      <td>[-0.04836796224117279, 0.1240588128566742, 0.0...</td>\n",
       "      <td>Frame_77_Face_83</td>\n",
       "      <td>77</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>28</td>\n",
       "      <td>[-0.08704757690429688, 0.023927267640829086, 0...</td>\n",
       "      <td>Frame_12_Face_12</td>\n",
       "      <td>12</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>29</td>\n",
       "      <td>[-0.062066659331321716, 0.1035008430480957, 0....</td>\n",
       "      <td>Frame_78_Face_84</td>\n",
       "      <td>78</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>29</td>\n",
       "      <td>[-0.0775049477815628, 0.13100804388523102, -0....</td>\n",
       "      <td>Frame_65_Face_70</td>\n",
       "      <td>65</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>25</td>\n",
       "      <td>[-0.0999695360660553, 0.12688468396663666, 0.0...</td>\n",
       "      <td>Frame_58_Face_63</td>\n",
       "      <td>58</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>30</td>\n",
       "      <td>[-0.08958522975444794, 0.13939359784126282, -0...</td>\n",
       "      <td>Frame_60_Face_65</td>\n",
       "      <td>60</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>27</td>\n",
       "      <td>[-0.10961633920669556, 0.022819139063358307, 0...</td>\n",
       "      <td>Frame_11_Face_11</td>\n",
       "      <td>11</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>22</td>\n",
       "      <td>[-0.10575947165489197, 0.06279823184013367, 0....</td>\n",
       "      <td>Frame_35_Face_32</td>\n",
       "      <td>35</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>31</td>\n",
       "      <td>[-0.07783067226409912, 0.04184330254793167, 0....</td>\n",
       "      <td>Frame_13_Face_13</td>\n",
       "      <td>13</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>32</td>\n",
       "      <td>[-0.193378284573555, 0.14120681583881378, 0.06...</td>\n",
       "      <td>Frame_39_Face_39</td>\n",
       "      <td>39</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63</th>\n",
       "      <td>28</td>\n",
       "      <td>[-0.11467906832695007, 0.1554570347070694, 0.0...</td>\n",
       "      <td>Frame_48_Face_55</td>\n",
       "      <td>48</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64</th>\n",
       "      <td>23</td>\n",
       "      <td>[-0.033346518874168396, 0.11731400340795517, 0...</td>\n",
       "      <td>Frame_30_Face_26</td>\n",
       "      <td>30</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65</th>\n",
       "      <td>29</td>\n",
       "      <td>[-0.12036427855491638, 0.09642638266086578, 0....</td>\n",
       "      <td>Frame_82_Face_85</td>\n",
       "      <td>82</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66</th>\n",
       "      <td>31</td>\n",
       "      <td>[-0.11149895191192627, 0.02655147761106491, 0....</td>\n",
       "      <td>Frame_14_Face_14</td>\n",
       "      <td>14</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>27</td>\n",
       "      <td>[-0.12408119440078735, 0.09628231823444366, 0....</td>\n",
       "      <td>Frame_38_Face_38</td>\n",
       "      <td>38</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68</th>\n",
       "      <td>20</td>\n",
       "      <td>[-0.061595797538757324, 0.1583673059940338, 0....</td>\n",
       "      <td>Frame_35_Face_31</td>\n",
       "      <td>35</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>33</td>\n",
       "      <td>[-0.043177515268325806, 0.1088676005601883, 0....</td>\n",
       "      <td>Frame_44_Face_46</td>\n",
       "      <td>44</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>23</td>\n",
       "      <td>[-0.08629979193210602, 0.09727758169174194, 0....</td>\n",
       "      <td>Frame_19_Face_20</td>\n",
       "      <td>19</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>25</td>\n",
       "      <td>[-0.1036529541015625, 0.11459621787071228, 0.1...</td>\n",
       "      <td>Frame_42_Face_43</td>\n",
       "      <td>42</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72</th>\n",
       "      <td>26</td>\n",
       "      <td>[-0.14419744908809662, 0.11870944499969482, 0....</td>\n",
       "      <td>Frame_70_Face_76</td>\n",
       "      <td>70</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73</th>\n",
       "      <td>30</td>\n",
       "      <td>[-0.050245970487594604, 0.06642231345176697, 0...</td>\n",
       "      <td>Frame_60_Face_66</td>\n",
       "      <td>60</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74</th>\n",
       "      <td>32</td>\n",
       "      <td>[-0.08343146741390228, 0.1408044695854187, 0.0...</td>\n",
       "      <td>Frame_63_Face_68</td>\n",
       "      <td>63</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>34</td>\n",
       "      <td>[-0.051602497696876526, 0.11674018204212189, 0...</td>\n",
       "      <td>Frame_76_Face_82</td>\n",
       "      <td>76</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>33</td>\n",
       "      <td>[-0.03690621256828308, 0.11352904140949249, 0....</td>\n",
       "      <td>Frame_47_Face_52</td>\n",
       "      <td>47</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77</th>\n",
       "      <td>31</td>\n",
       "      <td>[-0.08132253587245941, 0.07046635448932648, 0....</td>\n",
       "      <td>Frame_34_Face_30</td>\n",
       "      <td>34</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78</th>\n",
       "      <td>22</td>\n",
       "      <td>[-0.10782048106193542, 0.13687551021575928, 0....</td>\n",
       "      <td>Frame_46_Face_51</td>\n",
       "      <td>46</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79</th>\n",
       "      <td>24</td>\n",
       "      <td>[-0.0315370112657547, 0.113944873213768, 0.023...</td>\n",
       "      <td>Frame_21_Face_23</td>\n",
       "      <td>21</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80</th>\n",
       "      <td>21</td>\n",
       "      <td>[-0.04929281771183014, 0.14030632376670837, 0....</td>\n",
       "      <td>Frame_37_Face_35</td>\n",
       "      <td>37</td>\n",
       "      <td>M</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81</th>\n",
       "      <td>29</td>\n",
       "      <td>[-0.07108056545257568, 0.13075107336044312, 0....</td>\n",
       "      <td>Frame_71_Face_77</td>\n",
       "      <td>71</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82</th>\n",
       "      <td>27</td>\n",
       "      <td>[-0.0996004045009613, 0.12354321777820587, 0.0...</td>\n",
       "      <td>Frame_67_Face_73</td>\n",
       "      <td>67</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>83</th>\n",
       "      <td>34</td>\n",
       "      <td>[-0.0439348965883255, 0.11301878839731216, 0.0...</td>\n",
       "      <td>Frame_49_Face_56</td>\n",
       "      <td>49</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>31</td>\n",
       "      <td>[-0.041364505887031555, 0.10071046650409698, 0...</td>\n",
       "      <td>Frame_42_Face_42</td>\n",
       "      <td>42</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85</th>\n",
       "      <td>27</td>\n",
       "      <td>[-0.04843258857727051, 0.10660571604967117, 0....</td>\n",
       "      <td>Frame_74_Face_80</td>\n",
       "      <td>74</td>\n",
       "      <td>H</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>86 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    age                                        face_encode          filename  \\\n",
       "0    28  [-0.06104235351085663, 0.087809719145298, 0.00...  Frame_67_Face_74   \n",
       "1    31  [-0.047167032957077026, 0.07096441090106964, 0...  Frame_18_Face_18   \n",
       "2    26  [-0.03251942992210388, 0.12871268391609192, 0....  Frame_31_Face_27   \n",
       "3    30  [-0.08731923997402191, 0.059576183557510376, 0...    Frame_1_Face_1   \n",
       "4    31  [-0.13991206884384155, 0.0679367259144783, 0.0...  Frame_69_Face_75   \n",
       "5    27  [-0.020048364996910095, 0.08193252980709076, 0...  Frame_75_Face_81   \n",
       "6    32  [-0.058422237634658813, 0.10472062230110168, 0...  Frame_45_Face_48   \n",
       "7    37  [-0.08477601408958435, 0.13571122288703918, 0....  Frame_59_Face_64   \n",
       "8    25  [-0.06735698878765106, 0.10298667848110199, 0....  Frame_41_Face_40   \n",
       "9    31  [-0.08778735995292664, 0.09631244093179703, 0....  Frame_72_Face_78   \n",
       "10   29  [-0.0733812004327774, 0.0293489508330822, 0.04...  Frame_15_Face_15   \n",
       "11   28  [-0.04642969369888306, 0.08232296258211136, 0....  Frame_21_Face_22   \n",
       "12   25  [-0.04680822789669037, 0.07841528952121735, 0....  Frame_19_Face_19   \n",
       "13   30  [-0.11061060428619385, 0.12125469744205475, 0....  Frame_62_Face_67   \n",
       "14   28  [-0.12499348819255829, 0.1069939061999321, 0.0...  Frame_33_Face_29   \n",
       "15   25  [-0.11667513847351074, 0.11083956807851791, 0....  Frame_43_Face_45   \n",
       "16   22  [-0.08945988118648529, 0.1424393206834793, 0.0...  Frame_33_Face_28   \n",
       "17   23  [-0.12405925989151001, 0.09737150371074677, 0....  Frame_37_Face_36   \n",
       "18   32  [-0.05823671817779541, 0.11110067367553711, 0....  Frame_46_Face_50   \n",
       "19   28  [-0.06141555309295654, 0.05424799025058746, 0....    Frame_5_Face_5   \n",
       "20   30  [-0.07218991219997406, 0.0726621076464653, 0.0...    Frame_3_Face_3   \n",
       "21   27  [-0.07461020350456238, 0.08418428897857666, 0....  Frame_53_Face_59   \n",
       "22   27  [-0.07565182447433472, 0.07065467536449432, 0....  Frame_83_Face_86   \n",
       "23   25  [-0.11129900813102722, 0.15234169363975525, 0....  Frame_44_Face_47   \n",
       "24   33  [-0.058597758412361145, 0.0858473852276802, 0....  Frame_43_Face_44   \n",
       "25   24  [-0.10831055045127869, 0.14204099774360657, 0....  Frame_47_Face_53   \n",
       "26   29  [-0.0750880241394043, 0.06742691993713379, 0.0...  Frame_24_Face_25   \n",
       "27   30  [-0.04836796224117279, 0.1240588128566742, 0.0...  Frame_77_Face_83   \n",
       "28   28  [-0.08704757690429688, 0.023927267640829086, 0...  Frame_12_Face_12   \n",
       "29   29  [-0.062066659331321716, 0.1035008430480957, 0....  Frame_78_Face_84   \n",
       "..  ...                                                ...               ...   \n",
       "56   29  [-0.0775049477815628, 0.13100804388523102, -0....  Frame_65_Face_70   \n",
       "57   25  [-0.0999695360660553, 0.12688468396663666, 0.0...  Frame_58_Face_63   \n",
       "58   30  [-0.08958522975444794, 0.13939359784126282, -0...  Frame_60_Face_65   \n",
       "59   27  [-0.10961633920669556, 0.022819139063358307, 0...  Frame_11_Face_11   \n",
       "60   22  [-0.10575947165489197, 0.06279823184013367, 0....  Frame_35_Face_32   \n",
       "61   31  [-0.07783067226409912, 0.04184330254793167, 0....  Frame_13_Face_13   \n",
       "62   32  [-0.193378284573555, 0.14120681583881378, 0.06...  Frame_39_Face_39   \n",
       "63   28  [-0.11467906832695007, 0.1554570347070694, 0.0...  Frame_48_Face_55   \n",
       "64   23  [-0.033346518874168396, 0.11731400340795517, 0...  Frame_30_Face_26   \n",
       "65   29  [-0.12036427855491638, 0.09642638266086578, 0....  Frame_82_Face_85   \n",
       "66   31  [-0.11149895191192627, 0.02655147761106491, 0....  Frame_14_Face_14   \n",
       "67   27  [-0.12408119440078735, 0.09628231823444366, 0....  Frame_38_Face_38   \n",
       "68   20  [-0.061595797538757324, 0.1583673059940338, 0....  Frame_35_Face_31   \n",
       "69   33  [-0.043177515268325806, 0.1088676005601883, 0....  Frame_44_Face_46   \n",
       "70   23  [-0.08629979193210602, 0.09727758169174194, 0....  Frame_19_Face_20   \n",
       "71   25  [-0.1036529541015625, 0.11459621787071228, 0.1...  Frame_42_Face_43   \n",
       "72   26  [-0.14419744908809662, 0.11870944499969482, 0....  Frame_70_Face_76   \n",
       "73   30  [-0.050245970487594604, 0.06642231345176697, 0...  Frame_60_Face_66   \n",
       "74   32  [-0.08343146741390228, 0.1408044695854187, 0.0...  Frame_63_Face_68   \n",
       "75   34  [-0.051602497696876526, 0.11674018204212189, 0...  Frame_76_Face_82   \n",
       "76   33  [-0.03690621256828308, 0.11352904140949249, 0....  Frame_47_Face_52   \n",
       "77   31  [-0.08132253587245941, 0.07046635448932648, 0....  Frame_34_Face_30   \n",
       "78   22  [-0.10782048106193542, 0.13687551021575928, 0....  Frame_46_Face_51   \n",
       "79   24  [-0.0315370112657547, 0.113944873213768, 0.023...  Frame_21_Face_23   \n",
       "80   21  [-0.04929281771183014, 0.14030632376670837, 0....  Frame_37_Face_35   \n",
       "81   29  [-0.07108056545257568, 0.13075107336044312, 0....  Frame_71_Face_77   \n",
       "82   27  [-0.0996004045009613, 0.12354321777820587, 0.0...  Frame_67_Face_73   \n",
       "83   34  [-0.0439348965883255, 0.11301878839731216, 0.0...  Frame_49_Face_56   \n",
       "84   31  [-0.041364505887031555, 0.10071046650409698, 0...  Frame_42_Face_42   \n",
       "85   27  [-0.04843258857727051, 0.10660571604967117, 0....  Frame_74_Face_80   \n",
       "\n",
       "    frame gender  \n",
       "0      67      M  \n",
       "1      18      H  \n",
       "2      31      H  \n",
       "3       1      H  \n",
       "4      69      H  \n",
       "5      75      H  \n",
       "6      45      H  \n",
       "7      59      H  \n",
       "8      41      H  \n",
       "9      72      M  \n",
       "10     15      H  \n",
       "11     21      H  \n",
       "12     19      H  \n",
       "13     62      H  \n",
       "14     33      M  \n",
       "15     43      M  \n",
       "16     33      M  \n",
       "17     37      M  \n",
       "18     46      H  \n",
       "19      5      H  \n",
       "20      3      H  \n",
       "21     53      H  \n",
       "22     83      H  \n",
       "23     44      H  \n",
       "24     43      H  \n",
       "25     47      M  \n",
       "26     24      H  \n",
       "27     77      H  \n",
       "28     12      H  \n",
       "29     78      H  \n",
       "..    ...    ...  \n",
       "56     65      H  \n",
       "57     58      H  \n",
       "58     60      H  \n",
       "59     11      H  \n",
       "60     35      M  \n",
       "61     13      H  \n",
       "62     39      H  \n",
       "63     48      H  \n",
       "64     30      H  \n",
       "65     82      H  \n",
       "66     14      H  \n",
       "67     38      M  \n",
       "68     35      M  \n",
       "69     44      H  \n",
       "70     19      M  \n",
       "71     42      M  \n",
       "72     70      H  \n",
       "73     60      M  \n",
       "74     63      H  \n",
       "75     76      H  \n",
       "76     47      H  \n",
       "77     34      H  \n",
       "78     46      M  \n",
       "79     21      M  \n",
       "80     37      M  \n",
       "81     71      H  \n",
       "82     67      H  \n",
       "83     49      H  \n",
       "84     42      H  \n",
       "85     74      H  \n",
       "\n",
       "[86 rows x 5 columns]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Ler dos dados\n",
    "df = pd.DataFrame(json_lst)\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Kmeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "nclust = 5\n",
    "X      = [x for x in df.face_encode.values]\n",
    "\n",
    "kmeans = KMeans(n_clusters=nclust).fit(X)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([4, 3, 2, 2, 4, 0, 3, 0, 3, 0, 2, 3, 3, 0, 1, 1, 4, 1, 3, 2, 2, 2, 2,\n",
       "       1, 3, 1, 2, 0, 2, 0, 4, 2, 1, 0, 4, 3, 0, 3, 1, 0, 1, 0, 2, 3, 2, 2,\n",
       "       0, 1, 3, 2, 3, 2, 4, 2, 2, 2, 0, 0, 0, 2, 1, 2, 2, 1, 2, 2, 2, 1, 4,\n",
       "       3, 1, 1, 2, 4, 0, 0, 3, 2, 1, 1, 4, 0, 0, 3, 3, 0], dtype=int32)"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['label'] = kmeans.labels_\n",
    "kmeans.labels_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "k: 1\n",
      "k: 2\n",
      "k: 3\n",
      "k: 4\n",
      "k: 5\n",
      "k: 6\n",
      "k: 7\n",
      "k: 8\n",
      "k: 9\n",
      "k: 10\n",
      "k: 11\n",
      "k: 12\n",
      "k: 13\n",
      "k: 14\n",
      "k: 15\n",
      "k: 16\n",
      "k: 17\n",
      "k: 18\n",
      "k: 19\n",
      "k: 20\n",
      "k: 21\n",
      "k: 22\n",
      "k: 23\n",
      "k: 24\n",
      "k: 25\n",
      "k: 26\n",
      "k: 27\n",
      "k: 28\n",
      "k: 29\n",
      "k: 30\n",
      "k: 31\n",
      "k: 32\n",
      "k: 33\n",
      "k: 34\n",
      "k: 35\n",
      "k: 36\n",
      "k: 37\n",
      "k: 38\n",
      "k: 39\n",
      "k: 40\n",
      "k: 41\n",
      "k: 42\n",
      "k: 43\n",
      "k: 44\n",
      "k: 45\n",
      "k: 46\n",
      "k: 47\n",
      "k: 48\n",
      "k: 49\n",
      "k: 50\n",
      "k: 51\n",
      "k: 52\n",
      "k: 53\n",
      "k: 54\n",
      "k: 55\n",
      "k: 56\n",
      "k: 57\n",
      "k: 58\n",
      "k: 59\n",
      "k: 60\n",
      "k: 61\n",
      "k: 62\n",
      "k: 63\n",
      "k: 64\n",
      "k: 65\n",
      "k: 66\n",
      "k: 67\n",
      "k: 68\n",
      "k: 69\n",
      "k: 70\n",
      "k: 71\n",
      "k: 72\n",
      "k: 73\n",
      "k: 74\n",
      "k: 75\n",
      "k: 76\n",
      "k: 77\n",
      "k: 78\n",
      "k: 79\n",
      "k: 80\n",
      "k: 81\n",
      "k: 82\n",
      "k: 83\n",
      "k: 84\n",
      "k: 85\n",
      "k: 86\n",
      "k: 87\n",
      "k: 88\n",
      "k: 89\n",
      "k: 90\n",
      "k: 91\n",
      "k: 92\n",
      "k: 93\n",
      "k: 94\n",
      "k: 95\n",
      "k: 96\n",
      "k: 97\n",
      "k: 98\n",
      "k: 99\n"
     ]
    }
   ],
   "source": [
    "AvgSil = []\n",
    "clusters=range(1, 100)\n",
    "\n",
    "for k in clusters:\n",
    "    print(\"k:\",k)\n",
    "    # k-means\n",
    "    km = KMeans(n_clusters=nclust).fit(X)\n",
    "\n",
    "\n",
    "    # http://www.sthda.com/english/wiki/determining-the-optimal-number-of-clusters-3-must-known-methods-unsupervised-machine-learning   \n",
    "    # Average silhouette \n",
    "    avg_sil = silhouette_score(X, km.labels_)\n",
    "    AvgSil.append(avg_sil)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEKCAYAAAA4t9PUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f23280e98d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cluster: 0\n"
     ]
    }
   ],
   "source": [
    "# Var IntraGrupo\n",
    "plt.plot(clusters, AvgSil)\n",
    "plt.xlabel('Number of clusters')\n",
    "plt.ylabel('Average silhouette ')\n",
    "plt.show()\n",
    "\n",
    "print(\"Cluster:\", np.argmax(AvgSil))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Hierarchy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from scipy.cluster.hierarchy import ward, linkage, dendrogram \n",
    "\n",
    "dist = euclidean_distances(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(86, 86)"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dist.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/marlesson/anaconda3/lib/python3.6/site-packages/ipykernel/__main__.py:1: ClusterWarning: scipy.cluster: The symmetric non-negative hollow observation matrix looks suspiciously like an uncondensed distance matrix\n",
      "  if __name__ == '__main__':\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABDAAAAWYCAYAAABArDYhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCgwAAACgewoMAAAAoHsKDAAAAKB7/wPxvqZXH7QN/wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2309538358>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "linkage_matrix = linkage(dist, method='single')\n",
    "fig, ax = plt.subplots(figsize=(15, 20)) # set size\n",
    "ax = dendrogram(linkage_matrix, orientation=\"right\");\n",
    "\n",
    "plt.tight_layout() #show plot with tight layout\n",
    "\n",
    "#uncomment below to save figure\n",
    "#plt.savefig('ward_clusters.png', dpi=200) #save figure as ward_clusters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['g',\n",
       " 'g',\n",
       " 'r',\n",
       " 'r',\n",
       " 'r',\n",
       " 'r',\n",
       " 'r',\n",
       " 'r',\n",
       " 'r',\n",
       " 'r',\n",
       " 'c',\n",
       " 'c',\n",
       " 'c',\n",
       " 'c',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'm',\n",
       " 'b',\n",
       " 'b',\n",
       " 'b',\n",
       " 'b',\n",
       " 'y',\n",
       " 'y',\n",
       " 'y',\n",
       " 'y',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'k',\n",
       " 'b',\n",
       " 'b']"
      ]
     },
     "execution_count": 150,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ax['color_list']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.70098085])"
      ]
     },
     "execution_count": 138,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "face_recognition.face_distance(np.array([X[0]]), np.array(X[5]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.04716703,  0.07096441,  0.04107501, -0.0532998 , -0.06277665,\n",
       "        0.05742974, -0.06883474, -0.05386267,  0.23739003, -0.12374143,\n",
       "        0.20478944,  0.13719702, -0.12069285, -0.14703622, -0.01558149,\n",
       "        0.09186341, -0.12282016, -0.1711532 , -0.05092923, -0.02057138,\n",
       "       -0.01396421,  0.06725986,  0.02774967,  0.01380976, -0.12573192,\n",
       "       -0.30720061, -0.09241381, -0.05750314,  0.10996753, -0.10845552,\n",
       "        0.02633239,  0.06455113, -0.1937429 ,  0.01232423,  0.03160873,\n",
       "        0.11749883, -0.01982366, -0.00529113,  0.19283423,  0.0649192 ,\n",
       "       -0.17473668, -0.02070265, -0.03438266,  0.22951946,  0.19048122,\n",
       "       -0.01966125,  0.08182379, -0.06699146,  0.12743023, -0.21978915,\n",
       "        0.11473655,  0.16287214,  0.10947964,  0.05128021,  0.08433042,\n",
       "       -0.11668159, -0.01894376,  0.14755769, -0.155138  ,  0.07481098,\n",
       "        0.0391167 , -0.0587515 , -0.08060058,  0.01416403,  0.14110257,\n",
       "        0.14170043, -0.07461582, -0.27929085,  0.16860071, -0.18787491,\n",
       "       -0.05707891,  0.20150769, -0.09461631, -0.09328155, -0.26467612,\n",
       "        0.01876538,  0.43581566,  0.13634713, -0.11660397,  0.02013928,\n",
       "       -0.08519328, -0.01503047,  0.0243585 ,  0.05439731, -0.08736563,\n",
       "       -0.01698587, -0.08473713,  0.04499613,  0.18978895,  0.08872892,\n",
       "       -0.08683714,  0.11418964,  0.01121394,  0.05120221,  0.00528876,\n",
       "        0.03079861, -0.01132895, -0.01528607, -0.05470763, -0.05176469,\n",
       "       -0.01482663,  0.00547733,  0.01460771,  0.12383357, -0.11201937,\n",
       "        0.20223998, -0.04355051,  0.00851799,  0.03366506,  0.06198161,\n",
       "       -0.06542836, -0.00781629,  0.12271132, -0.2839621 ,  0.20143893,\n",
       "        0.09034406, -0.00095869,  0.10368065,  0.03533736,  0.13014157,\n",
       "       -0.04054768, -0.06967524, -0.11668906, -0.08790936, -0.0583472 ,\n",
       "       -0.05456059,  0.08295979, -0.01844848])"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(X[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DBSCAN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.cluster import DBSCAN\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "dist = euclidean_distances(X)\n",
    "db   = DBSCAN(eps=0.6, min_samples=5)\n",
    "labels = db.fit_predict(dist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-1,  0,  1,  1, -1,  2,  3, -1,  3,  2,  1,  0,  0,  4, -1,  5, -1,\n",
       "       -1,  3,  1,  1,  1,  1,  5,  3,  5,  1,  2,  1,  2, -1,  1,  5,  2,\n",
       "       -1,  3,  4, -1, -1,  4,  5,  4,  1,  0,  1, -1,  4,  5, -1,  1,  0,\n",
       "        1, -1,  1,  1,  1,  4,  4,  4,  1, -1,  1, -1,  5,  1, -1,  1, -1,\n",
       "       -1,  3, -1,  5, -1, -1,  4,  2,  3,  1,  5, -1, -1,  2,  4,  3,  3,\n",
       "        2])"
      ]
     },
     "execution_count": 189,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "df['label'] = labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for file in df[df['label'] == 6].filename.values:\n",
    "    imgshow(file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}