Exploratory Data Analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import gc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dtypes = {'ip': np.uint32, 'app': np.uint16, 'device': np.uint8, 'os': np.uint8, 'channel': np.uint8, 'is_attributed': np.bool}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "path = \"C:/Users/Admin/Desktop/mnt/ssd/kaggle-talkingdata2/competition_files/\"\n",
    "\n",
    "train = pd.read_csv(path + \"train_sample.csv\", sep=',', dtype=dtypes, parse_dates=['click_time', 'attributed_time'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ip</th>\n",
       "      <th>app</th>\n",
       "      <th>device</th>\n",
       "      <th>os</th>\n",
       "      <th>channel</th>\n",
       "      <th>click_time</th>\n",
       "      <th>attributed_time</th>\n",
       "      <th>is_attributed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>87540</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>241</td>\n",
       "      <td>2017-11-07 09:30:38</td>\n",
       "      <td>NaT</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>105560</td>\n",
       "      <td>25</td>\n",
       "      <td>1</td>\n",
       "      <td>17</td>\n",
       "      <td>3</td>\n",
       "      <td>2017-11-07 13:40:27</td>\n",
       "      <td>NaT</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>101424</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>212</td>\n",
       "      <td>2017-11-07 18:05:24</td>\n",
       "      <td>NaT</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>94584</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>221</td>\n",
       "      <td>2017-11-07 04:58:08</td>\n",
       "      <td>NaT</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>68413</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>178</td>\n",
       "      <td>2017-11-09 09:00:09</td>\n",
       "      <td>NaT</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       ip  app  device  os  channel          click_time attributed_time  \\\n",
       "0   87540   12       1  13      241 2017-11-07 09:30:38             NaT   \n",
       "1  105560   25       1  17        3 2017-11-07 13:40:27             NaT   \n",
       "2  101424   12       1  19      212 2017-11-07 18:05:24             NaT   \n",
       "3   94584   13       1  13      221 2017-11-07 04:58:08             NaT   \n",
       "4   68413   12       1   1      178 2017-11-09 09:00:09             NaT   \n",
       "\n",
       "   is_attributed  \n",
       "0          False  \n",
       "1          False  \n",
       "2          False  \n",
       "3          False  \n",
       "4          False  "
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 100000 entries, 0 to 99999\n",
      "Data columns (total 8 columns):\n",
      "ip                 100000 non-null uint32\n",
      "app                100000 non-null uint16\n",
      "device             100000 non-null uint8\n",
      "os                 100000 non-null uint8\n",
      "channel            100000 non-null uint8\n",
      "click_time         100000 non-null datetime64[ns]\n",
      "attributed_time    227 non-null datetime64[ns]\n",
      "is_attributed      100000 non-null bool\n",
      "dtypes: bool(1), datetime64[ns](2), uint16(1), uint32(1), uint8(3)\n",
      "memory usage: 2.5 MB\n"
     ]
    }
   ],
   "source": [
    "train.info()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train.drop(\"attributed_time\", axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7EAAAHRCAYAAAC8SD4qAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134d6ee8f28>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n=plt.figure(figsize=(16,8))\n",
    "cols = [\"ip\", \"app\", \"device\", \"os\", \"channel\"]\n",
    "uniques = [train[col].nunique() for col in cols]\n",
    "\n",
    "ax = sns.barplot(cols, uniques, log=True)\n",
    "ax.set_ylabel(\"unique count\")\n",
    "for p, uniq in zip(ax.patches, uniques):\n",
    "    ax.text(p.get_x() + p.get_width() / 2,\n",
    "           p.get_height(),\n",
    "           uniq,\n",
    "           ha=\"center\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7wAAAJMCAYAAADZpqqaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134d87564e0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "m=plt.figure(figsize=(16, 10))\n",
    "\n",
    "for i, col in enumerate([\"app\",\"device\",\"os\",\"channel\"]):\n",
    "    counts = pd.value_counts(train[col].values, ascending=False)[1:10]\n",
    "    indx = counts.index\n",
    "    counts.reset_index(drop=True, inplace=True)\n",
    "    \n",
    "    plt.subplot(2,2,i+1)\n",
    "    ax = sns.barplot(counts.index, counts.values, log=True, palette=\"Blues_d\")\n",
    "    ax.set_xticklabels(indx)\n",
    "    ax.set_xlabel(col)\n",
    "    ax.set_ylabel(\"count\")\n",
    "    \n",
    "    for p, prop in zip(ax.patches, [round(100*val/len(train),2) for val in counts.values]):\n",
    "        ax.text(p.get_x() + p.get_width() / 2,\n",
    "               p.get_height(),\n",
    "               prop,\n",
    "               ha=\"center\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "c=plt.figure(figsize=(16,8))\n",
    "\n",
    "ip_uniq = pd.value_counts(train[\"ip\"].values, ascending=False)[1:10]\n",
    "ip_indx = ip_uniq.index\n",
    "ip_uniq.reset_index(drop=True, inplace=True)\n",
    "\n",
    "ax = sns.barplot(ip_uniq.index, ip_uniq.values, log=True, palette=\"Blues_d\")\n",
    "ax.set_xticklabels(ip_indx)\n",
    "ax.set_xlabel(\"ip\")\n",
    "ax.set_ylabel(\"count\")\n",
    "\n",
    "for p, uniq in zip(ax.patches, [round(100*val/len(train), 3) for val in ip_uniq.values]):\n",
    "    ax.text(p.get_x() + p.get_width() / 2,\n",
    "           p.get_height(),\n",
    "           uniq,\n",
    "           ha=\"center\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7wAAAHfCAYAAABzvNTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134d4af3f60>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "c\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "z=plt.figure(figsize=(8,5))\n",
    "\n",
    "temp = train.groupby(\"is_attributed\").size()\n",
    "indx = temp.index\n",
    "temp.reset_index(drop=True, inplace=True)\n",
    "    \n",
    "ax = sns.barplot(temp.index, temp.values, log=True)\n",
    "ax.set_xticklabels(indx)\n",
    "ax.set_xlabel(\"is_attributed\")\n",
    "ax.set_ylabel(\"count\")\n",
    "\n",
    "for p, count in zip(ax.patches, temp.values):\n",
    "    ax.text(p.get_x() + p.get_width() / 2,\n",
    "           p.get_height(),\n",
    "           count,\n",
    "           ha='center')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfIAAAE8CAYAAAA7czMUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134d8da1208>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "z"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "\n",
    "corr=train[cols].corr()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ip</th>\n",
       "      <th>app</th>\n",
       "      <th>device</th>\n",
       "      <th>os</th>\n",
       "      <th>channel</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ip</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.010400</td>\n",
       "      <td>0.000177</td>\n",
       "      <td>0.000054</td>\n",
       "      <td>0.001838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>app</th>\n",
       "      <td>0.010400</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.203279</td>\n",
       "      <td>0.186648</td>\n",
       "      <td>-0.032750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>device</th>\n",
       "      <td>0.000177</td>\n",
       "      <td>0.203279</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.564018</td>\n",
       "      <td>-0.030028</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>os</th>\n",
       "      <td>0.000054</td>\n",
       "      <td>0.186648</td>\n",
       "      <td>0.564018</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.017930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>channel</th>\n",
       "      <td>0.001838</td>\n",
       "      <td>-0.032750</td>\n",
       "      <td>-0.030028</td>\n",
       "      <td>-0.017930</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               ip       app    device        os   channel\n",
       "ip       1.000000  0.010400  0.000177  0.000054  0.001838\n",
       "app      0.010400  1.000000  0.203279  0.186648 -0.032750\n",
       "device   0.000177  0.203279  1.000000  0.564018 -0.030028\n",
       "os       0.000054  0.186648  0.564018  1.000000 -0.017930\n",
       "channel  0.001838 -0.032750 -0.030028 -0.017930  1.000000"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAU4AAAD9CAYAAADTep4GAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDk4zs5z+P6vU25fcgGCyAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134d6211438>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "cax = ax.matshow(corr, vmin=-1, vmax=1)\n",
    "fig.colorbar(cax)\n",
    "ticks = np.arange(0,5,1)\n",
    "ax.set_xticks(ticks)\n",
    "ax.set_yticks(ticks)\n",
    "ax.set_xticklabels(train[cols])\n",
    "ax.set_yticklabels(train[cols])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ip</th>\n",
       "      <th>app</th>\n",
       "      <th>device</th>\n",
       "      <th>os</th>\n",
       "      <th>channel</th>\n",
       "      <th>click_time</th>\n",
       "      <th>is_attributed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>87540</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>497</td>\n",
       "      <td>11/7/2017 9:30</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>105560</td>\n",
       "      <td>25</td>\n",
       "      <td>1</td>\n",
       "      <td>17</td>\n",
       "      <td>259</td>\n",
       "      <td>11/7/2017 13:40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>101424</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>212</td>\n",
       "      <td>11/7/2017 18:05</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>94584</td>\n",
       "      <td>13</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>477</td>\n",
       "      <td>11/7/2017 4:58</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>68413</td>\n",
       "      <td>12</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>178</td>\n",
       "      <td>11/9/2017 9:00</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       ip  app  device  os  channel       click_time  is_attributed\n",
       "0   87540   12       1  13      497   11/7/2017 9:30              0\n",
       "1  105560   25       1  17      259  11/7/2017 13:40              0\n",
       "2  101424   12       1  19      212  11/7/2017 18:05              0\n",
       "3   94584   13       1  13      477   11/7/2017 4:58              0\n",
       "4   68413   12       1   1      178   11/9/2017 9:00              0"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train2 = pd.read_csv(path + \"train_sample.csv\", sep=',')\n",
    "train2.drop(\"attributed_time\", axis=1, inplace=True)\n",
    "train2.head()"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# extract day and hour from click_time column\n",
    "# add day and hour as new features\n",
    "# as we \n",
    "day, hour = [], []\n",
    "for time in train2[\"click_time\"]:\n",
    "    temp = time.split(\"/\")[1]\n",
    "    temp1 = time.split(\"/\")[-1]\n",
    "    temp2 = temp1.split(\":\")[0]\n",
    "    day_tp, hour_tp = temp,temp2.split(\" \")[-1]\n",
    "    day.append(int(day_tp))\n",
    "    hour.append(int(hour_tp))\n",
    "\n",
    "train2[\"day\"], train2[\"hour\"] = day, hour"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x21a933f4160>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "downloads = train2.groupby(\"hour\").apply(lambda x: len(x[x.is_attributed == 1]))\n",
    "clicks = train2.groupby(\"hour\").apply(lambda x: len(x[x.is_attributed == 0]))\n",
    "\n",
    "plt.figure(figsize=(16,8))\n",
    "\n",
    "plt.subplot(121)\n",
    "plt.bar(clicks.index, clicks.values)\n",
    "plt.xlabel(\"hour\")\n",
    "plt.ylabel(\"total clicks\")\n",
    "\n",
    "plt.subplot(122)\n",
    "plt.bar(downloads.index, downloads.values)\n",
    "plt.xlabel(\"hour\")\n",
    "plt.ylabel(\"total downloads\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x21a92cc5eb8>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_dayHour_data(is_attributed, click):\n",
    "    temp = train2.groupby([\"day\",\"hour\"]).apply(lambda x: len(x[x.is_attributed == is_attributed]))\n",
    "    days = temp.index.get_level_values(\"day\")\n",
    "    hours = temp.index.get_level_values(\"hour\")\n",
    "    \n",
    "    dayHour = []\n",
    "    for day, hour in zip(days, hours):\n",
    "        \n",
    "        dayHour.append(\"/\".join([str(day),str(hour) + 'H']))\n",
    "    \n",
    "    return pd.DataFrame({\"dayHour\":dayHour, click:temp.values}, index=np.arange(len(dayHour)))\n",
    "\n",
    "clickDownload = get_dayHour_data(0,\"clicks\").merge(get_dayHour_data(1,\"downloads\"), how=\"outer\", on=\"dayHour\")\n",
    "\n",
    "plt.figure(figsize=(16,8))\n",
    "\n",
    "plt.plot(range(len(clickDownload)), clickDownload.clicks, color='red')\n",
    "plt.plot(range(len(clickDownload)), clickDownload.downloads, color='blue')\n",
    "plt.yscale(\"log\")\n",
    "plt.legend()\n",
    "\n",
    "indx = [i for i in range(0, len(clickDownload), 5)]\n",
    "plt.xticks([i for i in range(0, len(clickDownload), 5)], \n",
    "           clickDownload.dayHour.values[indx],\n",
    "            rotation=45)\n",
    "\n",
    "plt.xlabel(\"day-hour\")\n",
    "plt.ylabel(\"count for clicks and downloads\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7YAAAHfCAYAAABknkQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAKTbEFAACg0BRbAAAACu3/A8bMhu7UCmACAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x21a8efdcf60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7EAAAHvCAYAAACCOFj3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRESywSSWiIiIiIiIZOP/AVtF1A4uHC8MAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x21a95bd4c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.show()"
   ]
  },
  {
   "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}