01_snapshots.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "toc-hr-collapsed": false
   },
   "source": [
    "# Snapshots"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.patches as patches\n",
    "import os\n",
    "import itertools\n",
    "\n",
    "from multiprocessing import Pool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%run config.py"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Valid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Devices:      3953\n",
      "Snapshots: 8830112\n"
     ]
    }
   ],
   "source": [
    "valid = pd.read_pickle(PATH_VALID_DEVICES, compression='gzip')\n",
    "\n",
    "print('Devices:   %7d' % len(valid))\n",
    "print('Snapshots: %7d' % valid.Count.sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "toc-hr-collapsed": false
   },
   "source": [
    "### Infos"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 0. SDK levels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "21     1.92\n",
       "22     5.97\n",
       "23    21.35\n",
       "24    17.51\n",
       "25     5.62\n",
       "26    30.20\n",
       "27    13.86\n",
       "28     3.57\n",
       "Name: SDK, dtype: float64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid.SDK.astype(int) \\\n",
    "    .value_counts(dropna=False, normalize=True) \\\n",
    "    .apply(lambda x: np.round(x * 100, 2)) \\\n",
    "    .sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7faf69d2b4e0>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "valid.SDK.astype(int).value_counts().plot(kind='bar')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1. Counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count     3953.000000\n",
       "mean      2233.774855\n",
       "std       1725.236381\n",
       "min        673.000000\n",
       "25%       1040.000000\n",
       "50%       1631.000000\n",
       "75%       2656.000000\n",
       "max      14257.000000\n",
       "Name: Count, dtype: float64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid.Count.astype(int).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7faf57c702e8>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "valid.Count.astype(int).hist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2. Manufacturers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "samsung          63.22\n",
       "lge               5.59\n",
       "generalmobile     5.21\n",
       "motorola          3.79\n",
       "sony              3.11\n",
       "oneplus           2.88\n",
       "lenovo            2.18\n",
       "xiaomi            2.12\n",
       "vestel            1.75\n",
       "htc               1.69\n",
       "Name: Manufacturer, dtype: float64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "manufacturers = valid.Manufacturer.apply(lambda x: x.lower()).value_counts(dropna=False, normalize=True)\n",
    "manufacturers.apply(lambda x: np.round(x * 100, 2)).head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "74"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(manufacturers)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3. Locales"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tr_TR          0.568935\n",
       "en_US          0.091323\n",
       "en_GB          0.051100\n",
       "es_US          0.050594\n",
       "de_DE          0.047053\n",
       "es_ES          0.027321\n",
       "en_IN          0.019985\n",
       "az_AZ_#Latn    0.017961\n",
       "it_IT          0.013913\n",
       "ru_RU          0.013408\n",
       "Name: Locale, dtype: float64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "locales = valid.Locale.value_counts(dropna=False, normalize=True)\n",
    "locales.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "76"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(locales)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tr    57.22\n",
       "en    17.76\n",
       "es     9.89\n",
       "de     5.34\n",
       "az     2.88\n",
       "it     1.39\n",
       "ru     1.39\n",
       "pt     1.16\n",
       "in     0.56\n",
       "fr     0.51\n",
       "Name: Locale, dtype: float64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "locales_split = valid.Locale.apply(lambda x: x.split('_')[0]).value_counts(dropna=False, normalize=True)\n",
    "locales_split.apply(lambda x: np.round(x * 100, 2)).head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "31"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(locales_split)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4. Timezones"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Europe/Istanbul                   0.516064\n",
       "Asia/Baku                         0.067797\n",
       "Europe/Berlin                     0.043005\n",
       "Asia/Kolkata                      0.035416\n",
       "Asia/Calcutta                     0.021503\n",
       "America/Mexico_City               0.018720\n",
       "America/New_York                  0.015684\n",
       "America/Argentina/Buenos_Aires    0.014925\n",
       "Europe/Madrid                     0.014419\n",
       "America/Bogota                    0.014166\n",
       "Name: Timezone, dtype: float64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "timezones = valid.Timezone.value_counts(dropna=False, normalize=True)\n",
    "timezones.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "158"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(timezones)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Europe       64.99\n",
       "Asia         18.54\n",
       "America      13.84\n",
       "Africa        1.82\n",
       "Australia     0.30\n",
       "Etc           0.23\n",
       "Pacific       0.18\n",
       "Atlantic      0.08\n",
       "GMT           0.03\n",
       "Name: Timezone, dtype: float64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "timezones_split = valid.Timezone.apply(lambda x: x.split('/')[0]).value_counts(dropna=False, normalize=True)\n",
    "timezones_split.apply(lambda x: np.round(x * 100, 2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(timezones_split)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5. Diffs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    3953.000000\n",
       "mean       17.744708\n",
       "std         3.563459\n",
       "min         5.270006\n",
       "25%        15.689997\n",
       "50%        16.757560\n",
       "75%        18.556238\n",
       "max        52.924250\n",
       "Name: DiffMean, dtype: float64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid.DiffMean.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7faf57bbf128>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "valid.DiffMin.plot(kind='hist', title='Min')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7faf57b587b8>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "valid.DiffMax.plot(kind='hist', title='Max')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7faf57bb5710>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "valid.DiffMean.plot(kind='hist', title='Mean')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7faf57a5d5f8>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "valid.DiffSD.plot(kind='hist', title='SD')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7faf579da9b0>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEICAYAAABI7RO5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAGRxJREFUeJzt3X+0XWV95/H3x4CCigLlymACJmqqg22N9Ap2aTsUB+RHK7iWpTCtRobVOGtgVadOR3DNFNTShV0q1hmlxRIFqqapoGQUi0GwjrMGIdHIz1qugiUxkmhAQBwY4Dt/nOfKabz35my45/7Ifb/WOuvs/d3P3ufZnHA/a/84z05VIUnSoJ422x2QJM0vBockqRODQ5LUicEhSerE4JAkdWJwSJI6MTikGZakkry4Tf9lkv82232Suoi/45Aml+Qu4PnA86vqh331bwIrgGVVdVfHbRawvKrGprGr0ozxiEPatTuBU8dnkvwy8MzZ6440uwwOadcuA97cN78SuHR8Jskzkrw/yT8nuaedftq7b/kfJ9ma5PtJ/n3/hpN8Ismftun9knw+yfYk97bpJX1tv5LkvUn+d5IHknwpyQFD22tpEgaHtGvXA89J8q+TLAJOAf6mb/n5wC/SO3X1YmAx8CcASY4F/jNwNLAc+LdTfM7TgI8DLwAOAX4K/I+d2vw74DTgecDT27alGWVwSIMZP+o4Grgd2NLqAVYB/6mqdlTVA8Cf0QsXgJOBj1fVLVX1E+DcyT6gqn5UVZdX1UNtO+cB/2anZh+vqn+qqp8Ca+mFlTSj9pjtDkjzxGXAV4Fl9J2mAkboXe/YmGS8FmBRm34+sLGv/fcm+4AkzwQuAI4F9mvlfZIsqqrH2vwP+lZ5CHh25z2RniKPOKQBVNX36F0kPx64om/RD+mdUnpZVe3bXs+tqvE/6FuBg/vaHzLFx7wDeAlwRFU9B/iNVs/kq0gzz+CQBnc6cFQ75TTuceBjwAVJngeQZHGS17Xla4G3JDm0HVGcM8X296EXQvcl2X8XbaVZY3BIA6qq71TVhgkWvRMYA65Pcj9wDb0jB6rqi8CHgGtbm2un+IgPAXvTO4q5Hvj76eu9NH38AaAkqROPOCRJnRgckqRODA5JUicGhySpk93yB4AHHHBALV26dLa7IUnzysaNG39YVSO7ajf04Ghj+2wAtlTVbyVZBqwBfoHeL2rfVFWPJHkGvV/k/irwI+B3x4erTnI2vXvoHwP+sKqunuozly5dyoYNE901KUmaTJJJRzboNxOnqt5Gb2yfce8DLqiqFwP30gsE2vu9rX5Ba0eSQ+mN+/MyekMxfLSFkSRpFgw1ONqQ0CcAf93mAxwFfKY1uQQ4qU2f2OZpy1/b2p8IrKmqh6vqTno/ojp8mP2WJE1u2EccHwL+C71hGaB3euq+qnq0zW+mNwQ17f1ugLb8x639z+oTrPMzSVYl2ZBkw/bt26d7PyRJzdCCI8lvAduqauMuG0+DqrqoqkaranRkZJfXdiRJT9IwL46/Gnh9kuOBvYDnAH8B7Jtkj3ZUsYQnnmuwhd4oopuT7AE8l95F8vH6uP51JEkzbGhHHFV1dlUtqaql9C5uX1tVvwdcB7yxNVsJXNmm17V52vJrqzeQ1jrglPZ4zmX0nqJ2w7D6LUma2mz8juOdwJr2nOVvAhe3+sXAZUnGgB20J6hV1a1J1gK3AY8CZ/Q91EaSNMN2y9FxR0dHy99xSFI3STZW1eiu2jnkiCSpk91yyBFpLlt61hdm5XPvOv+EWflc7X484pAkdWJwSJI6MTgkSZ0YHJKkTgwOSVInBockqRODQ5LUicEhSerE4JAkdWJwSJI6MTgkSZ0YHJKkTgwOSVInBockqRODQ5LUicEhSepkaMGRZK8kNyT5VpJbk7y71T+R5M4km9prRasnyYeTjCW5KclhfdtameSO9lo5rD5LknZtmE8AfBg4qqoeTLIn8LUkX2zL/riqPrNT++OA5e11BHAhcESS/YFzgFGggI1J1lXVvUPsuyRpEkM74qieB9vsnu1VU6xyInBpW+96YN8kBwGvA9ZX1Y4WFuuBY4fVb0nS1IZ6jSPJoiSbgG30/vh/vS06r52OuiDJM1ptMXB33+qbW22y+s6ftSrJhiQbtm/fPu37IknqGWpwVNVjVbUCWAIcnuSXgLOBlwKvBPYH3jlNn3VRVY1W1ejIyMh0bFKSNIEZuauqqu4DrgOOraqt7XTUw8DHgcNbsy3AwX2rLWm1yeqSpFkwzLuqRpLs26b3Bo4G/rFdtyBJgJOAW9oq64A3t7urXgX8uKq2AlcDxyTZL8l+wDGtJkmaBcO8q+og4JIki+gF1Nqq+nySa5OMAAE2Af+htb8KOB4YAx4CTgOoqh1J3gvc2Nq9p6p2DLHfkqQpDC04quom4BUT1I+apH0BZ0yybDWwelo7KEl6UvzluCSpE4NDktSJwSFJ6sTgkCR1YnBIkjoxOCRJnRgckqRODA5JUicGhySpE4NDktSJwSFJ6sTgkCR1YnBIkjoxOCRJnRgckqRODA5JUicGhySpk2E+c3yvJDck+VaSW5O8u9WXJfl6krEkf5vk6a3+jDY/1pYv7dvW2a3+7SSvG1afJUm7NswjjoeBo6rq5cAK4NgkrwLeB1xQVS8G7gVOb+1PB+5t9QtaO5IcCpwCvAw4Fvhoe465JGkWDC04qufBNrtnexVwFPCZVr8EOKlNn9jmactfmyStvqaqHq6qO4Ex4PBh9VuSNLWhXuNIsijJJmAbsB74DnBfVT3ammwGFrfpxcDdAG35j4Ff6K9PsE7/Z61KsiHJhu3btw9jdyRJDDk4quqxqloBLKF3lPDSIX7WRVU1WlWjIyMjw/oYSVrwZuSuqqq6D7gO+DVg3yR7tEVLgC1tegtwMEBb/lzgR/31CdaRJM2wYd5VNZJk3za9N3A0cDu9AHlja7YSuLJNr2vztOXXVlW1+intrqtlwHLghmH1W5I0tT123eRJOwi4pN0B9TRgbVV9PsltwJokfwp8E7i4tb8YuCzJGLCD3p1UVNWtSdYCtwGPAmdU1WND7LckaQpDC46qugl4xQT17zLBXVFV9X+B35lkW+cB5013HyVJ3fnLcUlSJwaHJKkTg0OS1InBIUnqxOCQJHVicEiSOjE4JEmdGBySpE4MDklSJwaHJKkTg0OS1InBIUnqxOCQJHVicEiSOjE4JEmdGBySpE4MDklSJwaHJKmToQVHkoOTXJfktiS3Jnlbq5+bZEuSTe11fN86ZycZS/LtJK/rqx/bamNJzhpWnyVJuza0Z44DjwLvqKpvJNkH2JhkfVt2QVW9v79xkkOBU4CXAc8Hrknyi23xR4Cjgc3AjUnWVdVtQ+y7JGkSQwuOqtoKbG3TDyS5HVg8xSonAmuq6mHgziRjwOFt2VhVfRcgyZrW1uCQpFkwI9c4kiwFXgF8vZXOTHJTktVJ9mu1xcDdfattbrXJ6jt/xqokG5Js2L59+zTvgSRp3EDBkeSXn+wHJHk2cDnw9qq6H7gQeBGwgt4RyQee7Lb7VdVFVTVaVaMjIyPTsUlJ0gQGPeL4aJIbkvzHJM8ddONJ9qQXGp+sqisAquqeqnqsqh4HPsYTp6O2AAf3rb6k1SarS5JmwUDBUVW/DvwevT/gG5N8KsnRU62TJMDFwO1V9cG++kF9zd4A3NKm1wGnJHlGkmXAcuAG4EZgeZJlSZ5O7wL6uoH2TpI07Qa+OF5VdyT5r8AG4MPAK1o4vGv8aGInrwbeBNycZFOrvQs4NckKoIC7gLe27d+aZC29i96PAmdU1WMASc4ErgYWAaur6tbOeypJmhYDBUeSXwFOA04A1gO/3W6zfT7wf4CfC46q+hqQCTZ31WSfU1XnAedNUL9qqvUkSTNn0COO/w78Nb2ji5+OF6vq++0oRJK0QAwaHCcAP+07dfQ0YK+qeqiqLhta7yRJc86gd1VdA+zdN//MVpMkLTCDBsdeVfXg+EybfuZwuiRJmssGDY6fJDlsfCbJrwI/naK9JGk3Neg1jrcDf5fk+/TulPpXwO8OrVeSpDlroOCoqhuTvBR4SSt9u6r+3/C6JUmaq7qMjvtKYGlb57AkVNWlQ+mVJGnOGvQHgJfRG5hwE/BYKxdgcEjSAjPoEccocGhV1TA7I0ma+wa9q+oWehfEJUkL3KBHHAcAtyW5AXh4vFhVrx9KryRJc9agwXHuMDshSZo/Br0d9x+SvABYXlXXJHkmvSHOJUkLzKCPjv0D4DPAX7XSYuBzw+qUJGnuGvTi+Bn0Hsx0P/Qe6gQ8b1idkiTNXYMGx8NV9cj4TJI96P2OQ5K0wAwaHP+Q5F3A3u1Z438H/M/hdUuSNFcNGhxnAduBm+k9I/wqYMon/yU5OMl1SW5LcmuSt7X6/knWJ7mjve/X6kny4SRjSW7aaTTela39HUlWPpkdlSRNj0Hvqnoc+Fh7DepR4B3t2eT7ABuTrAfeAny5qs5Pcha9UHoncBywvL2OAC4EjkiyP3AOvV+vV9vOuqq6t0NfJEnTZNCxqu5kgmsaVfXCydapqq3A1jb9QJLb6d2NdSJwZGt2CfAVesFxInBpG9bk+iT7JjmotV1fVTtaX9YDxwKfHqTvkqTp1WWsqnF7Ab8D7D/ohyRZCrwC+DpwYAsVgB8AB7bpxcDdfattbrXJ6jt/xipgFcAhhxwyaNckSR0NdI2jqn7U99pSVR8CThhk3STPBi4H3l5V9++03WKa7s6qqouqarSqRkdGRqZjk5KkCQx6quqwvtmn0TsC2eW6SfakFxqfrKorWvmeJAdV1dZ2Kmpbq28BDu5bfUmrbeGJU1vj9a8M0m9J0vQb9FTVB/qmHwXuAk6eaoUkAS4Gbq+qD/YtWgesBM5v71f21c9MsobexfEft3C5Gviz8buvgGOAswfstyRpmg16V9VvPoltvxp4E3Bzkk2t9i56gbE2yenA93gigK4CjgfGgIeA09pn70jyXuDG1u494xfKJUkzb9BTVX801fKdjijGa18DMskqr52gfdEb2mSi7a8GVu+6p5KkYetyV9Ur6Z1OAvht4AbgjmF0SpI0dw0aHEuAw6rqAYAk5wJfqKrfH1bHJElz06BDjhwIPNI3/whP/P5CkrSADHrEcSlwQ5LPtvmT6P3qW5K0wAx6V9V5Sb4I/HornVZV3xxetyRJc9Wgp6oAngncX1V/AWxOsmxIfZIkzWGDPjr2HHoDEY7/8G5P4G+G1SlJ0tw16BHHG4DXAz8BqKrvA/sMq1OSpLlr0OB4pH9AwiTPGl6XJElz2aDBsTbJXwH7JvkD4Bq6PdRJkrSbGPSuqve3Z43fD7wE+JOqWj/UnkmS5qRBhkZfBFzTBjo0LCRpgdvlqaqqegx4PMlzZ6A/kqQ5btBfjj9Ib3j09bQ7qwCq6g+H0itJ0pw1aHBc0V6SpAVuyuBIckhV/XNVOS6VJAnY9TWOz41PJLl8yH2RJM0DuwqO/if4vbDLhpOsTrItyS19tXOTbEmyqb2O71t2dpKxJN9O8rq++rGtNpbkrC59kCRNv10FR00yPYhPAMdOUL+gqla011UASQ4FTgFe1tb5aJJF7VbgjwDHAYcCp7a2kqRZsquL4y9Pcj+9I4+92zRtvqrqOZOtWFVfTbJ0wH6cCKypqoeBO5OMAYe3ZWNV9V2AJGta29sG3K4kaZpNecRRVYuq6jlVtU9V7dGmx+cnDY1dODPJTe1U1n6tthi4u6/N5labrC5JmiVdnscxHS4EXgSsALYCH5iuDSdZlWRDkg3bt2+frs1KknYyo8FRVfdU1WNV9Ti9QRLHT0dtAQ7ua7qk1SarT7Tti6pqtKpGR0ZGpr/zkiRghoMjyUF9s28Axu+4WgeckuQZ7cmCy4EbgBuB5UmWJXk6vQvo62ayz5Kkf2nQX453luTTwJHAAUk2A+cARyZZQe8OrbuAtwJU1a1J1tK76P0ocEYbI4skZwJXA4uA1VV167D6LEnataEFR1WdOkH54inanwecN0H9KuCqaeyaJOkpmOmL45Kkec7gkCR1YnBIkjoxOCRJnRgckqRODA5JUicGhySpE4NDktSJwSFJ6sTgkCR1YnBIkjoxOCRJnRgckqRODA5JUicGhySpE4NDktSJwSFJ6sTgkCR1MrTgSLI6ybYkt/TV9k+yPskd7X2/Vk+SDycZS3JTksP61lnZ2t+RZOWw+itJGswwjzg+ARy7U+0s4MtVtRz4cpsHOA5Y3l6rgAuhFzTAOcARwOHAOeNhI0maHUMLjqr6KrBjp/KJwCVt+hLgpL76pdVzPbBvkoOA1wHrq2pHVd0LrOfnw0iSNINm+hrHgVW1tU3/ADiwTS8G7u5rt7nVJqv/nCSrkmxIsmH79u3T22tJ0s/M2sXxqiqgpnF7F1XVaFWNjoyMTNdmJUk7menguKedgqK9b2v1LcDBfe2WtNpkdUnSLJnp4FgHjN8ZtRK4sq/+5nZ31auAH7dTWlcDxyTZr10UP6bVJEmzZI9hbTjJp4EjgQOSbKZ3d9T5wNokpwPfA05uza8CjgfGgIeA0wCqakeS9wI3tnbvqaqdL7hLkmbQ0IKjqk6dZNFrJ2hbwBmTbGc1sHoauyZJegr85bgkqRODQ5LUicEhSerE4JAkdWJwSJI6MTgkSZ0YHJKkTgwOSVInBockqRODQ5LUicEhSerE4JAkdWJwSJI6MTgkSZ0YHJKkTgwOSVInBockqRODQ5LUyawER5K7ktycZFOSDa22f5L1Se5o7/u1epJ8OMlYkpuSHDYbfZYk9czmEcdvVtWKqhpt82cBX66q5cCX2zzAccDy9loFXDjjPZUk/cxcOlV1InBJm74EOKmvfmn1XA/sm+Sg2eigJGn2gqOALyXZmGRVqx1YVVvb9A+AA9v0YuDuvnU3t9q/kGRVkg1JNmzfvn1Y/ZakBW+PWfrc11TVliTPA9Yn+cf+hVVVSarLBqvqIuAigNHR0U7rSpIGNytHHFW1pb1vAz4LHA7cM34Kqr1va823AAf3rb6k1SRJs2DGgyPJs5LsMz4NHAPcAqwDVrZmK4Er2/Q64M3t7qpXAT/uO6UlSZphs3Gq6kDgs0nGP/9TVfX3SW4E1iY5HfgecHJrfxVwPDAGPAScNvNdliSNm/HgqKrvAi+foP4j4LUT1As4Ywa6JkkawFy6HVeSNA8YHJKkTgwOSVInBockqRODQ5LUicEhSerE4JAkdWJwSJI6MTgkSZ0YHJKkTgwOSVInBockqRODQ5LUicEhSerE4JAkdWJwSJI6MTgkSZ0YHJKkTuZNcCQ5Nsm3k4wlOWu2+yNJC9WMP3P8yUiyCPgIcDSwGbgxybqqum12e6b5aulZX5jtLkjz1rwIDuBwYKyqvguQZA1wImBwzHP+AZ85s/nf+q7zT5i1z9b0my/BsRi4u29+M3BEf4Mkq4BVbfbBJN8eYLsHAD+clh7OHe7T/LCg9invm+GeTK+F9F29YJCV50tw7FJVXQRc1GWdJBuqanRIXZoV7tP84D7NH7vjfj3VfZovF8e3AAf3zS9pNUnSDJsvwXEjsDzJsiRPB04B1s1ynyRpQZoXp6qq6tEkZwJXA4uA1VV16zRsutOprXnCfZof3Kf5Y3fcr6e0T6mq6eqIJGkBmC+nqiRJc4TBIUnqZEEGx+46fEmSu5LcnGRTkg2z3Z8nI8nqJNuS3NJX2z/J+iR3tPf9ZrOPXU2yT+cm2dK+q01Jjp/NPnaV5OAk1yW5LcmtSd7W6vP2u5pin+btd5VkryQ3JPlW26d3t/qyJF9vfwP/tt10NPh2F9o1jjZ8yT/RN3wJcOruMHxJkruA0aqatz9WSvIbwIPApVX1S63258COqjq/Bf1+VfXO2exnF5Ps07nAg1X1/tns25OV5CDgoKr6RpJ9gI3AScBbmKff1RT7dDLz9LtKEuBZVfVgkj2BrwFvA/4IuKKq1iT5S+BbVXXhoNtdiEccPxu+pKoeAcaHL9EcUFVfBXbsVD4RuKRNX0Lvf+Z5Y5J9mteqamtVfaNNPwDcTm+Eh3n7XU2xT/NW9TzYZvdsrwKOAj7T6p2/p4UYHBMNXzKv/3H0KeBLSTa2IVh2FwdW1dY2/QPgwNnszDQ6M8lN7VTWvDmls7MkS4FXAF9nN/mudtonmMffVZJFSTYB24D1wHeA+6rq0dak89/AhRgcu7PXVNVhwHHAGe0UyW6leudWd4fzqxcCLwJWAFuBD8xud56cJM8GLgfeXlX39y+br9/VBPs0r7+rqnqsqlbQG3HjcOClT3WbCzE4dtvhS6pqS3vfBnyW3j+S3cE97fzz+HnobbPcn6esqu5p/0M/DnyMefhdtXPmlwOfrKorWnlef1cT7dPu8F0BVNV9wHXArwH7Jhn/AXjnv4ELMTh2y+FLkjyrXdAjybOAY4Bbpl5r3lgHrGzTK4ErZ7Ev02L8j2vzBubZd9Uuul4M3F5VH+xbNG+/q8n2aT5/V0lGkuzbpvemd1PQ7fQC5I2tWefvacHdVQXQbqf7EE8MX3LeLHfpKUvyQnpHGdAbSuZT83G/knwaOJLesM/3AOcAnwPWAocA3wNOrqp5c7F5kn06kt6pjwLuAt7ad21gzkvyGuB/ATcDj7fyu+hdE5iX39UU+3Qq8/S7SvIr9C5+L6J3oLC2qt7T/l6sAfYHvgn8flU9PPB2F2JwSJKevIV4qkqS9BQYHJKkTgwOSVInBockqRODQ5LUicEhSerE4JAkdfL/AfEbB/vS8V7qAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "valid.DiffMedian.plot(kind='hist', title='Median')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 6. Durations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count                        3953\n",
       "mean      27 days 04:55:51.149479\n",
       "std       20 days 21:31:15.777385\n",
       "min        7 days 00:33:47.496000\n",
       "25%       12 days 16:45:25.760000\n",
       "50%       19 days 23:20:19.785000\n",
       "75%       32 days 02:55:53.001000\n",
       "max      109 days 19:25:24.862000\n",
       "Name: Duration, dtype: object"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid.Duration.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7faf57959588>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "valid.Duration.dt.days.hist(bins=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "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": "3.6.7"
  },
  "toc-autonumbering": false,
  "toc-showtags": false
 },
 "nbformat": 4,
 "nbformat_minor": 2
}