COMPAS_Python_solutions.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%bash\n",
    "# Some set-up specific to google colab.\n",
    "!(stat -t /usr/local/lib/*/dist-packages/google/colab > /dev/null 2>&1) && exit \n",
    "git clone https://github.com/SIDN-IAP/compas-python.git compas-python 2>> install.log\n",
    "cp compas-python/compas-scores-two-years.csv ."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# COMPAS analysis\n",
    "\n",
    "We recreate the first section of the [Propublica COMPAS analysis](https://github.com/propublica/compas-analysis) in Python. \n",
    "\n",
    "What follows are the calculations performed for ProPublica's analaysis of the COMPAS Recidivism Risk Scores. It might be helpful to open [the methodology](https://www.propublica.org/article/how-we-analyzed-the-compas-recidivism-algorithm/) in another tab to understand the following.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading the Data\n",
    "\n",
    "We select fields for severity of charge, number of priors, demographics, age, sex, compas scores, and whether each person was accused of a crime within two years.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Num rows: 7214\n"
     ]
    }
   ],
   "source": [
    "raw_data = pd.read_csv('./compas-scores-two-years.csv')\n",
    "print('Num rows: %d' %len(raw_data))\n",
    "\n",
    "df = raw_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "However not all of the rows are useable for the first round of analysis.\n",
    "\n",
    "There are a number of reasons remove rows because of missing data:\n",
    "\n",
    " - If the charge date of a defendants Compas scored crime was not within 30 days from when the person was arrested, we assume that because of data quality reasons, that we do not have the right offense.\n",
    " - We coded the recidivist flag -- is_recid -- to be -1 if we could not find a compas case at all.\n",
    " - In a similar vein, ordinary traffic offenses -- those with a df of 'O' -- will not result in Jail time are removed (only two of them).\n",
    " - We filtered the underlying data from Broward county to include only those rows representing people who had either recidivated in two years, or had at least two years outside of a correctional facility.\n",
    " - We remove rows where there is no score_text ('N/A')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Num rows filtered: 6172\n"
     ]
    }
   ],
   "source": [
    "df = df[((df['days_b_screening_arrest'] <=30) & \n",
    "      (df['days_b_screening_arrest'] >= -30) &\n",
    "      (df['is_recid'] != -1) &\n",
    "      (df['c_charge_degree'] != 'O') & \n",
    "      (df['score_text'] != 'N/A')\n",
    "     )]\n",
    "\n",
    "print('Num rows filtered: %d' % len(df))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Higher COMPAS scores are slightly correlated with a longer length of stay.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from datetime import datetime\n",
    "from scipy.stats import pearsonr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def date_from_str(s):\n",
    "    return datetime.strptime(s, '%Y-%m-%d %H:%M:%S')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Correlation btw stay length and COMPAS scores: 0.207\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/afs/csail.mit.edu/u/d/davidbau/.conda/envs/p3t13/lib/python3.6/site-packages/ipykernel_launcher.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    }
   ],
   "source": [
    "df['length_of_stay'] = (df['c_jail_out'].apply(date_from_str) - df['c_jail_in'].apply(date_from_str)).dt.total_seconds()\n",
    "stay_score_corr = pearsonr(df['length_of_stay'], df['decile_score'])[0]\n",
    "print('Correlation btw stay length and COMPAS scores: %.3f' % stay_score_corr)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After filtering we have the following demographic breakdown:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "25 - 45            3532\n",
       "Less than 25       1347\n",
       "Greater than 45    1293\n",
       "Name: age_cat, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['age_cat'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "African-American    3175\n",
       "Caucasian           2103\n",
       "Hispanic             509\n",
       "Other                343\n",
       "Asian                 31\n",
       "Native American       11\n",
       "Name: race, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['race'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Black defendants: 51.44%\n",
      "White defendants: 34.07%\n",
      "Hispanic defendants: 8.25%\n",
      "Asian defendants: 0.50%\n",
      "Native American defendants: 0.18%\n"
     ]
    }
   ],
   "source": [
    "print(\"Black defendants: %.2f%%\" %            (3175 / 6172 * 100))\n",
    "print(\"White defendants: %.2f%%\" %            (2103 / 6172 * 100))\n",
    "print(\"Hispanic defendants: %.2f%%\" %         (509  / 6172 * 100))\n",
    "print(\"Asian defendants: %.2f%%\" %            (31   / 6172 * 100))\n",
    "print(\"Native American defendants: %.2f%%\" %  (11   / 6172 * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Low       3421\n",
       "Medium    1607\n",
       "High      1144\n",
       "Name: score_text, dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['score_text'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>race</th>\n",
       "      <th>African-American</th>\n",
       "      <th>Asian</th>\n",
       "      <th>Caucasian</th>\n",
       "      <th>Hispanic</th>\n",
       "      <th>Native American</th>\n",
       "      <th>Other</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sex</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>Female</th>\n",
       "      <td>549</td>\n",
       "      <td>2</td>\n",
       "      <td>482</td>\n",
       "      <td>82</td>\n",
       "      <td>2</td>\n",
       "      <td>58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Male</th>\n",
       "      <td>2626</td>\n",
       "      <td>29</td>\n",
       "      <td>1621</td>\n",
       "      <td>427</td>\n",
       "      <td>9</td>\n",
       "      <td>285</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "race    African-American  Asian  Caucasian  Hispanic  Native American  Other\n",
       "sex                                                                         \n",
       "Female               549      2        482        82                2     58\n",
       "Male                2626     29       1621       427                9    285"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.crosstab(df.sex, df.race)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Male      4997\n",
       "Female    1175\n",
       "Name: sex, dtype: int64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['sex'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Men: 80.96%\n",
      "Women: 19.04%\n"
     ]
    }
   ],
   "source": [
    "print(\"Men: %.2f%%\" %   (4997 / 6172 * 100))\n",
    "print(\"Women: %.2f%%\" % (1175 / 6172 * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Num with two-year recidivism: 6172\n"
     ]
    }
   ],
   "source": [
    "print('Num with two-year recidivism: %d' % len(df['two_year_recid'] == 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Num with two-year recidivism: 45.51%\n"
     ]
    }
   ],
   "source": [
    "print('Num with two-year recidivism: %.2f%%' % (len(df[df['two_year_recid'] == 1]) / len(df) * 100.))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Judges are often presented with two sets of scores from the Compas system -- one that classifies people into High, Medium and Low risk, and a corresponding decile score. There is a clear downward trend in the decile scores as those scores increase for white defendants.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "from matplotlib import pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Count')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 277,
       "width": 388
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "black_df = df[df['race'] == 'African-American']\n",
    "black_df['decile_score'].hist(edgecolor='white', bins=np.arange(13)-0.5,grid=False)\n",
    "plt.ylim([0,700])\n",
    "plt.xlim([0,11])\n",
    "plt.title('Black defendant decile scores')\n",
    "plt.xlabel('Score decile')\n",
    "plt.ylabel('Count')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Score decile')"
      ]
     },
     "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": {
      "image/png": {
       "height": 277,
       "width": 375
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "black_df = df[df['race'] == 'Caucasian']\n",
    "black_df['decile_score'].hist(edgecolor='white', bins=np.arange(13)-0.5,grid=False)\n",
    "plt.ylim([0,700])\n",
    "plt.xlim([0,11])\n",
    "plt.title('White defendants decile scores')\n",
    "plt.xlabel('Score decile')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>race</th>\n",
       "      <th>African-American</th>\n",
       "      <th>Asian</th>\n",
       "      <th>Caucasian</th>\n",
       "      <th>Hispanic</th>\n",
       "      <th>Native American</th>\n",
       "      <th>Other</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>decile_score</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</th>\n",
       "      <td>365</td>\n",
       "      <td>15</td>\n",
       "      <td>605</td>\n",
       "      <td>159</td>\n",
       "      <td>0</td>\n",
       "      <td>142</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>346</td>\n",
       "      <td>4</td>\n",
       "      <td>321</td>\n",
       "      <td>89</td>\n",
       "      <td>2</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>298</td>\n",
       "      <td>5</td>\n",
       "      <td>238</td>\n",
       "      <td>73</td>\n",
       "      <td>1</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>337</td>\n",
       "      <td>0</td>\n",
       "      <td>243</td>\n",
       "      <td>47</td>\n",
       "      <td>0</td>\n",
       "      <td>39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>323</td>\n",
       "      <td>1</td>\n",
       "      <td>200</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>318</td>\n",
       "      <td>2</td>\n",
       "      <td>160</td>\n",
       "      <td>27</td>\n",
       "      <td>2</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>343</td>\n",
       "      <td>1</td>\n",
       "      <td>113</td>\n",
       "      <td>28</td>\n",
       "      <td>2</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>301</td>\n",
       "      <td>2</td>\n",
       "      <td>96</td>\n",
       "      <td>14</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>317</td>\n",
       "      <td>0</td>\n",
       "      <td>77</td>\n",
       "      <td>17</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>227</td>\n",
       "      <td>1</td>\n",
       "      <td>50</td>\n",
       "      <td>16</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "race          African-American  Asian  Caucasian  Hispanic  Native American  \\\n",
       "decile_score                                                                  \n",
       "1                          365     15        605       159                0   \n",
       "2                          346      4        321        89                2   \n",
       "3                          298      5        238        73                1   \n",
       "4                          337      0        243        47                0   \n",
       "5                          323      1        200        39                0   \n",
       "6                          318      2        160        27                2   \n",
       "7                          343      1        113        28                2   \n",
       "8                          301      2         96        14                0   \n",
       "9                          317      0         77        17                2   \n",
       "10                         227      1         50        16                2   \n",
       "\n",
       "race          Other  \n",
       "decile_score         \n",
       "1               142  \n",
       "2                60  \n",
       "3                32  \n",
       "4                39  \n",
       "5                19  \n",
       "6                20  \n",
       "7                 9  \n",
       "8                 7  \n",
       "9                 7  \n",
       "10                8  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.crosstab(df.decile_score, df.race)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Racial Bias in Compas\n",
    "\n",
    "After filtering out bad rows, our first question is whether there is a significant difference in Compas scores between races. To do so we need to change some variables into factors, and run a logistic regression, comparing low scores to high scores.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "lr = LogisticRegression(solver='lbfgs')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_crime = pd.get_dummies(df['c_charge_degree'],prefix='crimefactor',drop_first=True)\n",
    "df_age = pd.get_dummies(df['age_cat'],prefix='age')\n",
    "df_race = pd.get_dummies(df['race'],prefix='race')\n",
    "df_gender = pd.get_dummies(df['sex'],prefix='sex',drop_first=True)\n",
    "df_score = pd.get_dummies(df['score_text'] != 'Low',prefix='score_factor',drop_first=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_lr = pd.concat([df_crime, df_age,df_race,df_gender,\n",
    "                   df['priors_count'],df['two_year_recid']\n",
    "                  ],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
       "                   intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
       "                   multi_class='warn', n_jobs=None, penalty='l2',\n",
       "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
       "                   warm_start=False)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr.fit(df_lr,df_score.values.ravel())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "cols = np.concatenate([df_crime.columns.values, \n",
    "                df_age.columns.values,\n",
    "                df_race.columns.values,\n",
    "                df_gender.columns.values,\n",
    "                ['priors_count'],\n",
    "                ['two_year_recid']\n",
    "               ])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-1.335 (Intercept)\n",
      "-0.310 crimefactor_M\n",
      "0.018 age_25 - 45\n",
      "-1.328 age_Greater than 45\n",
      "1.319 age_Less than 25\n",
      "0.490 race_African-American\n",
      "-0.201 race_Asian\n",
      "0.013 race_Caucasian\n",
      "-0.409 race_Hispanic\n",
      "0.909 race_Native American\n",
      "-0.793 race_Other\n",
      "-0.219 sex_Male\n",
      "0.268 priors_count\n",
      "0.683 two_year_recid\n"
     ]
    }
   ],
   "source": [
    "assert len(cols) == len(lr.coef_[0])\n",
    "print('%.3f' % lr.intercept_, '(Intercept)') \n",
    "for i,j in zip(cols, lr.coef_[0]): print('%.3f' % j, i)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Black defendants are 44% more likely than white defendants to receive a higher score correcting for the seriousness of their crime, previous arrests, and future criminal behavior.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Black defendants: 1.4423\n"
     ]
    }
   ],
   "source": [
    "control = np.exp(-1.335) / (1 + np.exp(-1.335))\n",
    "black_odds = np.exp(0.490) / (1 - control + (control * np.exp(0.490)))\n",
    "print('Black defendants: %.4f' % black_odds)"
   ]
  }
 ],
 "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"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}