Fairness metrics.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "import matplotlib\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import pandas\n",
    "from scipy.stats import ttest_ind\n",
    "from IPython.display import display"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "adult_df = pandas.read_csv(\"./data/adult.csv\", index_col=0)\n",
    "compas_df = pandas.read_csv(\"./data/compas.csv\", index_col=0)\n",
    "bank_df = pandas.read_csv(\"./data/bank.csv\", index_col=0)\n",
    "german_df = pandas.read_csv(\"./data/german.csv\", index_col=0)\n",
    "\n",
    "'''\n",
    "\n",
    "adult_df = pandas.read_csv(\"./data/All_Data -adult.csv\", index_col=0)\n",
    "compas_df = pandas.read_csv(\"./data/All_Data -compas.csv\", index_col=0)\n",
    "bank_df = pandas.read_csv(\"./data/All_Data -bank.csv\", index_col=0)\n",
    "german_df = pandas.read_csv(\"./data/All_Data -german.csv\", index_col=0)\n",
    "\n",
    "\n",
    "for df in [adult_df, compas_df, bank_df, german_df]:\n",
    "    df[\"False Positive Rate Diff\"] = df[\"False Positive Rate - Unprivileged\"] - df[\"False Positive Rate - Privileged\"]\n",
    "    df[\"False Negative Rate Diff\"] = df[\"False Negative Rate - Unprivileged\"] - df[\"False Negative Rate - Privileged\"]\n",
    "    df[\"False Omission Rate Diff\"] = df[\"False Omission Rate - Unprivileged\"] - df[\"False Omission Rate - Privileged\"]\n",
    "    df[\"False Discovery Rate Diff\"] = df[\"False Discovery Rate - Unprivileged\"] - df[\"False Discovery Rate - Privileged\"]\n",
    "    df[\"Statistical Parity Diff\"] = df[\"Privileged base Rate\"] - df[\"Unprivileged base Rate\"] \n",
    "    df[\"Accuracy Diff\"] = df[\"Accuracy - Privileged\"] - df[\"Accuracy - Unprivileged\"] \n",
    "    df[\"F1 Score Diff\"] = df[\"F1 Score - Privileged\"] - df[\"F1 Score - Unprivileged\"] \n",
    "    df[\"base Rate\"] = (df[\"Num True Pos\"] + df[\"Num False Pos\"])/(df[\"Num True Pos\"] + df[\"Num True Neg\"] + df[\"Num False Pos\"] + df[\"Num False Neg\"])\n",
    "    for metric in df.columns:\n",
    "        if \"Num\" in metric:\n",
    "            df.drop(metric, axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Logistic Regression - 3', 'Logistic Regression - 11', 'Logistic Regression - 17']\n",
      "['OP - 3', 'OP - 11', 'OP - 17', 'DIR - 3', 'DIR - 11', 'DIR - 17', 'EOP - 3', 'EOP - 11', 'EOP - 17', 'CEOP - 3', 'CEOP - 11', 'CEOP - 17', 'ROC - 3', 'ROC - 11', 'ROC - 17', 'GFC - 3', 'GFC - 11', 'GFC - 17', 'PR - 3', 'PR - 11', 'PR - 17', 'EGR - 3', 'EGR - 11', 'EGR - 17', 'GSR - 3', 'GSR - 11', 'GSR - 17']\n",
      "['OP + EOP - 3', 'OP + EOP - 11', 'OP + EOP - 17', 'OP + CEOP - 3', 'OP + CEOP - 11', 'OP + CEOP - 17', 'OP + ROC - 3', 'OP + ROC - 11', 'OP + ROC - 17', 'OP + GFC - 3', 'OP + GFC - 11', 'OP + GFC - 17', 'OP + PR - 3', 'OP + PR - 11', 'OP + PR - 17', 'OP + EGR - 3', 'OP + EGR - 11', 'OP + EGR - 17', 'OP + GSR - 3', 'OP + GSR - 11', 'OP + GSR - 17', 'DIR + EOP - 3', 'DIR + EOP - 11', 'DIR + EOP - 17', 'DIR + CEOP - 3', 'DIR + CEOP - 11', 'DIR + CEOP - 17', 'DIR + ROC - 3', 'DIR + ROC - 11', 'DIR + ROC - 17', 'DIR + GFC - 3', 'DIR + GFC - 11', 'DIR + GFC - 17', 'DIR + PR - 3', 'DIR + PR - 11', 'DIR + PR - 17', 'DIR + EGR - 3', 'DIR + EGR - 11', 'DIR + EGR - 17', 'DIR + GSR - 3', 'DIR + GSR - 11', 'DIR + GSR - 17', 'GFC + EOP - 3', 'GFC + EOP - 11', 'GFC + EOP - 17', 'GFC + CEOP - 3', 'GFC + CEOP - 11', 'GFC + CEOP - 17', 'GFC + ROC - 3', 'GFC + ROC - 11', 'GFC + ROC - 17', 'PR + EOP - 3', 'PR + EOP - 11', 'PR + EOP - 17', 'PR + CEOP - 3', 'PR + CEOP - 11', 'PR + CEOP - 17', 'PR + ROC - 3', 'PR + ROC - 11', 'PR + ROC - 17', 'EGR + EOP - 3', 'EGR + EOP - 11', 'EGR + EOP - 17', 'EGR + CEOP - 3', 'EGR + CEOP - 11', 'EGR + CEOP - 17', 'EGR + ROC - 3', 'EGR + ROC - 11', 'EGR + ROC - 17', 'GSR + EOP - 3', 'GSR + EOP - 11', 'GSR + EOP - 17', 'GSR + CEOP - 3', 'GSR + CEOP - 11', 'GSR + CEOP - 17', 'GSR + ROC - 3', 'GSR + ROC - 11', 'GSR + ROC - 17']\n",
      "['OP + GFC + EOP - 3', 'OP + GFC + EOP - 11', 'OP + GFC + EOP - 17', 'OP + GFC + CEOP - 3', 'OP + GFC + CEOP - 11', 'OP + GFC + CEOP - 17', 'OP + GFC + ROC - 3', 'OP + GFC + ROC - 11', 'OP + GFC + ROC - 17', 'OP + PR + EOP - 3', 'OP + PR + EOP - 11', 'OP + PR + EOP - 17', 'OP + PR + CEOP - 3', 'OP + PR + CEOP - 11', 'OP + PR + CEOP - 17', 'OP + PR + ROC - 3', 'OP + PR + ROC - 11', 'OP + PR + ROC - 17', 'OP + EGR + EOP - 3', 'OP + EGR + EOP - 11', 'OP + EGR + EOP - 17', 'OP + EGR + CEOP - 3', 'OP + EGR + CEOP - 11', 'OP + EGR + CEOP - 17', 'OP + EGR + ROC - 3', 'OP + EGR + ROC - 11', 'OP + EGR + ROC - 17', 'OP + GSR + EOP - 3', 'OP + GSR + EOP - 11', 'OP + GSR + EOP - 17', 'OP + GSR + CEOP - 3', 'OP + GSR + CEOP - 11', 'OP + GSR + CEOP - 17', 'OP + GSR + ROC - 3', 'OP + GSR + ROC - 11', 'OP + GSR + ROC - 17', 'DIR + GFC + EOP - 3', 'DIR + GFC + EOP - 11', 'DIR + GFC + EOP - 17', 'DIR + GFC + CEOP - 3', 'DIR + GFC + CEOP - 11', 'DIR + GFC + CEOP - 17', 'DIR + GFC + ROC - 3', 'DIR + GFC + ROC - 11', 'DIR + GFC + ROC - 17', 'DIR + PR + EOP - 3', 'DIR + PR + EOP - 11', 'DIR + PR + EOP - 17', 'DIR + PR + CEOP - 3', 'DIR + PR + CEOP - 11', 'DIR + PR + CEOP - 17', 'DIR + PR + ROC - 3', 'DIR + PR + ROC - 11', 'DIR + PR + ROC - 17', 'DIR + EGR + EOP - 3', 'DIR + EGR + EOP - 11', 'DIR + EGR + EOP - 17', 'DIR + EGR + CEOP - 3', 'DIR + EGR + CEOP - 11', 'DIR + EGR + CEOP - 17', 'DIR + EGR + ROC - 3', 'DIR + EGR + ROC - 11', 'DIR + EGR + ROC - 17', 'DIR + GSR + EOP - 3', 'DIR + GSR + EOP - 11', 'DIR + GSR + EOP - 17', 'DIR + GSR + CEOP - 3', 'DIR + GSR + CEOP - 11', 'DIR + GSR + CEOP - 17', 'DIR + GSR + ROC - 3', 'DIR + GSR + ROC - 11', 'DIR + GSR + ROC - 17']\n"
     ]
    }
   ],
   "source": [
    "def num_interventions(name):\n",
    "    if \"Logistic Regression\" in name:\n",
    "        return 0\n",
    "    if name.count('+') == 0:\n",
    "        return 1\n",
    "    if name.count('+') == 1:\n",
    "        return 2\n",
    "    if name.count('+') == 2:\n",
    "        return 3\n",
    "    \n",
    "no_int = []\n",
    "one_int = []\n",
    "two_int = []\n",
    "three_int = []\n",
    "for row in list(adult_df.index):\n",
    "    if num_interventions(row) == 0:\n",
    "        no_int.append(row)\n",
    "    elif num_interventions(row) == 1:\n",
    "        one_int.append(row)\n",
    "    elif num_interventions(row) == 2:\n",
    "        two_int.append(row)\n",
    "    elif num_interventions(row) == 3:\n",
    "        three_int.append(row)\n",
    "        \n",
    "print(no_int)\n",
    "print(one_int)\n",
    "print(two_int)\n",
    "print(three_int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['False Positive Rate Diff',\n",
       " 'False Negative Rate Diff',\n",
       " 'False Omission Rate Diff',\n",
       " 'False Discovery Rate Diff',\n",
       " 'Statistical Parity Diff',\n",
       " 'Accuracy Diff',\n",
       " 'F1 Score Diff']"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "col_names = []\n",
    "for c in adult_df.columns:\n",
    "    if \"Diff\" in c:\n",
    "        col_names.append(c)\n",
    "\n",
    "col_names #= ['False Positive Rate Diff']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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></th>\n",
       "      <th>False Positive Rate Diff</th>\n",
       "      <th>False Negative Rate Diff</th>\n",
       "      <th>False Omission Rate Diff</th>\n",
       "      <th>False Discovery Rate Diff</th>\n",
       "      <th>Statistical Parity Diff</th>\n",
       "      <th>Accuracy Diff</th>\n",
       "      <th>F1 Score Diff</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>OP - 3</th>\n",
       "      <td>-0.0387</td>\n",
       "      <td>0.1468</td>\n",
       "      <td>-0.1392</td>\n",
       "      <td>0.3015</td>\n",
       "      <td>0.1167</td>\n",
       "      <td>-0.1145</td>\n",
       "      <td>0.2041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP - 11</th>\n",
       "      <td>-0.0351</td>\n",
       "      <td>0.1628</td>\n",
       "      <td>-0.1363</td>\n",
       "      <td>0.3375</td>\n",
       "      <td>0.1134</td>\n",
       "      <td>-0.1101</td>\n",
       "      <td>0.2294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP - 17</th>\n",
       "      <td>-0.0399</td>\n",
       "      <td>0.1679</td>\n",
       "      <td>-0.1196</td>\n",
       "      <td>0.3013</td>\n",
       "      <td>0.1185</td>\n",
       "      <td>-0.0987</td>\n",
       "      <td>0.2213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP + EOP - 3</th>\n",
       "      <td>-0.0006</td>\n",
       "      <td>-0.0077</td>\n",
       "      <td>-0.1755</td>\n",
       "      <td>0.3059</td>\n",
       "      <td>0.0420</td>\n",
       "      <td>-0.1369</td>\n",
       "      <td>0.0719</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP + EOP - 11</th>\n",
       "      <td>0.0063</td>\n",
       "      <td>0.0145</td>\n",
       "      <td>-0.1698</td>\n",
       "      <td>0.3526</td>\n",
       "      <td>0.0386</td>\n",
       "      <td>-0.1240</td>\n",
       "      <td>0.1143</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GSR + CEOP - 11</th>\n",
       "      <td>0.1343</td>\n",
       "      <td>-0.4699</td>\n",
       "      <td>-0.2010</td>\n",
       "      <td>0.4328</td>\n",
       "      <td>-0.0986</td>\n",
       "      <td>-0.0453</td>\n",
       "      <td>-0.0002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GSR + CEOP - 17</th>\n",
       "      <td>-0.0728</td>\n",
       "      <td>0.5255</td>\n",
       "      <td>-0.0701</td>\n",
       "      <td>-0.2349</td>\n",
       "      <td>0.2136</td>\n",
       "      <td>-0.0802</td>\n",
       "      <td>0.6231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GSR + ROC - 3</th>\n",
       "      <td>0.0630</td>\n",
       "      <td>-0.1368</td>\n",
       "      <td>-0.0979</td>\n",
       "      <td>0.3297</td>\n",
       "      <td>0.0420</td>\n",
       "      <td>0.0378</td>\n",
       "      <td>0.2456</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GSR + ROC - 11</th>\n",
       "      <td>0.0229</td>\n",
       "      <td>-0.4365</td>\n",
       "      <td>-0.2328</td>\n",
       "      <td>0.2328</td>\n",
       "      <td>-0.0482</td>\n",
       "      <td>-0.2068</td>\n",
       "      <td>-0.4321</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GSR + ROC - 17</th>\n",
       "      <td>0.0408</td>\n",
       "      <td>-0.1781</td>\n",
       "      <td>-0.1118</td>\n",
       "      <td>0.2925</td>\n",
       "      <td>0.0508</td>\n",
       "      <td>0.0002</td>\n",
       "      <td>0.1796</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>180 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 False Positive Rate Diff  False Negative Rate Diff  \\\n",
       "OP - 3                            -0.0387                    0.1468   \n",
       "OP - 11                           -0.0351                    0.1628   \n",
       "OP - 17                           -0.0399                    0.1679   \n",
       "OP + EOP - 3                      -0.0006                   -0.0077   \n",
       "OP + EOP - 11                      0.0063                    0.0145   \n",
       "...                                   ...                       ...   \n",
       "GSR + CEOP - 11                    0.1343                   -0.4699   \n",
       "GSR + CEOP - 17                   -0.0728                    0.5255   \n",
       "GSR + ROC - 3                      0.0630                   -0.1368   \n",
       "GSR + ROC - 11                     0.0229                   -0.4365   \n",
       "GSR + ROC - 17                     0.0408                   -0.1781   \n",
       "\n",
       "                 False Omission Rate Diff  False Discovery Rate Diff  \\\n",
       "OP - 3                            -0.1392                     0.3015   \n",
       "OP - 11                           -0.1363                     0.3375   \n",
       "OP - 17                           -0.1196                     0.3013   \n",
       "OP + EOP - 3                      -0.1755                     0.3059   \n",
       "OP + EOP - 11                     -0.1698                     0.3526   \n",
       "...                                   ...                        ...   \n",
       "GSR + CEOP - 11                   -0.2010                     0.4328   \n",
       "GSR + CEOP - 17                   -0.0701                    -0.2349   \n",
       "GSR + ROC - 3                     -0.0979                     0.3297   \n",
       "GSR + ROC - 11                    -0.2328                     0.2328   \n",
       "GSR + ROC - 17                    -0.1118                     0.2925   \n",
       "\n",
       "                 Statistical Parity Diff  Accuracy Diff  F1 Score Diff  \n",
       "OP - 3                            0.1167        -0.1145         0.2041  \n",
       "OP - 11                           0.1134        -0.1101         0.2294  \n",
       "OP - 17                           0.1185        -0.0987         0.2213  \n",
       "OP + EOP - 3                      0.0420        -0.1369         0.0719  \n",
       "OP + EOP - 11                     0.0386        -0.1240         0.1143  \n",
       "...                                  ...            ...            ...  \n",
       "GSR + CEOP - 11                  -0.0986        -0.0453        -0.0002  \n",
       "GSR + CEOP - 17                   0.2136        -0.0802         0.6231  \n",
       "GSR + ROC - 3                     0.0420         0.0378         0.2456  \n",
       "GSR + ROC - 11                   -0.0482        -0.2068        -0.4321  \n",
       "GSR + ROC - 17                    0.0508         0.0002         0.1796  \n",
       "\n",
       "[180 rows x 7 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adult_df[col_names]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 12 0.023809523809523808 0.05323971374999499\n",
      "1 105 0.09251700680272107 0.12621568591356863\n",
      "2 291 0.10260186548846342 0.14907750342058756\n",
      "3 252 0.12414965986394555 0.16852188331544637\n"
     ]
    }
   ],
   "source": [
    "priv = {0:[], 1:[], 2:[], 3:[]}\n",
    "for dataset in [adult_df, german_df, compas_df, bank_df]:\n",
    "    d_sub = dataset[col_names]\n",
    "    d_sub = np.sum(abs(d_sub)<=0.01, axis=1)/len(col_names)\n",
    "    for inter, val in d_sub.items():\n",
    "        k = num_interventions(inter)\n",
    "        priv[k].append(val)\n",
    "\n",
    "y_mean = []\n",
    "for k in priv:\n",
    "    print(k, len(priv[k]), np.mean(priv[k]), np.std(priv[k]))\n",
    "    y_mean.append(round(np.mean(priv[k])*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# percent of times privileged group performed worse than unprivileged group\n",
    "x = [\"Zero\", \"One\", \"Two\", \"Three\"]\n",
    "fig, ax = plt.subplots(figsize=(7, 5))\n",
    "c = 'sandybrown'\n",
    "rects1 = ax.bar(x, y_mean, color=c, edgecolor=c, alpha=0.55, width=0.6, linewidth=2)\n",
    "\n",
    "plt.rcParams.update({'font.size': 17})\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "ax.spines['left'].set_visible(False)\n",
    "ax.spines['bottom'].set_color('#DDDDDD')\n",
    "ax.tick_params(bottom=False, left=False)\n",
    "ax.set_axisbelow(True)\n",
    "ax.yaxis.grid(True, color='#EEEEEE')\n",
    "ax.xaxis.grid(False)\n",
    "ax.bar_label(rects1, padding=3)\n",
    "ax.set_ylim([0, 14])\n",
    "\n",
    "ax.set_ylabel('% times disparity <1%', size=15)\n",
    "#ax.set_ylabel('Group Fairness')\n",
    "ax.set_xlabel('Number of Interventions', size=15)\n",
    "\n",
    "fig.tight_layout()\n",
    "#plt.show()\n",
    "fig.savefig('./figures/group_fair_inter.png', dpi=600)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3>Individual Bias</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([14, 16, 19, 19],\n",
       " [89, 90, 91, 91],\n",
       " [1.22, 1.03, 0.9, 1.03],\n",
       " [1.67, 0.7, 0.46, 0.61])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ind_bias = {0:{\"theil\":[], \"consis\":[]}, 1:{\"theil\":[], \"consis\":[]}, 2:{\"theil\":[], \"consis\":[]}, 3:{\"theil\":[], \"consis\":[]}}\n",
    "for dataset in [adult_df, german_df, compas_df, bank_df]:\n",
    "    d = dataset[[\"Theil Index\", \"Consistency\"]]\n",
    "    for index, row in d.iterrows():\n",
    "        k = num_interventions(index)\n",
    "        #print(index, k, row[\"Theil Index\"], row[\"Consistency\"])\n",
    "        ind_bias[k][\"theil\"].append(row[\"Theil Index\"])\n",
    "        ind_bias[k][\"consis\"].append(row[\"Consistency\"])\n",
    "\n",
    "theil_mean = [round(np.mean(ind_bias[k][\"theil\"])*100) for k in [0,1,2,3]]\n",
    "consis_mean = [round(np.mean(ind_bias[k][\"consis\"])*100) for k in [0,1,2,3]]\n",
    "\n",
    "theil_std = [round(np.std(ind_bias[k][\"theil\"])/(len(ind_bias[k][\"theil\"])**0.5)*100, 2) for k in [0,1,2,3]]\n",
    "consis_std = [round(np.std(ind_bias[k][\"consis\"])/(len(ind_bias[k][\"theil\"])**0.5)*100, 2) for k in [0,1,2,3]]\n",
    "\n",
    "theil_mean, consis_mean, theil_std, consis_std  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# percent of times privileged group performed worse than unprivileged group\n",
    "x = np.arange(4)\n",
    "x_lab = [\"Zero\", \"One\", \"Two\", \"Three\"]\n",
    "fig, ax = plt.subplots(figsize=(9, 5))\n",
    "\n",
    "width = 0.4 \n",
    "c1, c2 = 'royalblue', 'darksalmon'\n",
    "rects1 = ax.bar(x - width/2, theil_mean, label='Theil Index', yerr=theil_std, color=c1, edgecolor=c1, linewidth=2, alpha=0.65, width=0.33, hatch='.')\n",
    "rects2 = ax.bar(x + width/2, consis_mean, label='Consistency', yerr=consis_std, color=c2, edgecolor=c2, linewidth=2, alpha=0.4, width=0.33, hatch='x')\n",
    "\n",
    "plt.rcParams.update({'font.size': 17})\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "ax.spines['left'].set_visible(False)\n",
    "ax.spines['bottom'].set_color('#DDDDDD')\n",
    "ax.tick_params(bottom=False, left=False)\n",
    "ax.set_axisbelow(True)\n",
    "ax.yaxis.grid(True, color='#EEEEEE')\n",
    "ax.xaxis.grid(False)\n",
    "ax.bar_label(rects1, padding=3)\n",
    "ax.bar_label(rects2, padding=3)\n",
    "ax.set_ylim([0, 109])\n",
    "ax.set_xticks(x, x_lab)\n",
    "#ax.set_ylabel('% of times group fairness metric <0.01')\n",
    "ax.set_ylabel('Theil, Consistency (10^-2)', size=15)\n",
    "ax.set_xlabel('Number of Interventions', size=15)\n",
    "ax.legend(bbox_to_anchor=(0.5, 1.13), loc='upper center', ncol=2)\n",
    "fig.tight_layout()\n",
    "#plt.show()\n",
    "fig.savefig('./figures/ind_fair.png', dpi=600)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3>% Dominanace</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([56, 53, 52, 53], [42, 38, 38, 35])"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# percentage of times one group is dominant over the other\n",
    "priv_dom = {0:[], 1:[], 2:[], 3:[]}\n",
    "unpriv_dom = {0:[], 1:[], 2:[], 3:[]}\n",
    "thresh = 0.01\n",
    "for dataset in [adult_df, german_df, compas_df, bank_df]:\n",
    "#for dataset in [adult_df, compas_df]:\n",
    "    d_sub = dataset[col_names]\n",
    "    #d_sub = dataset[[\"False Positive Rate Diff\"]]\n",
    "    priv_dom_df = np.sum(d_sub>thresh, axis=1)/len(col_names)\n",
    "    unpriv_dom_df = np.sum(d_sub< -thresh, axis=1)/len(col_names)\n",
    "    for inter, val in priv_dom_df.items():\n",
    "        k = num_interventions(inter)\n",
    "        priv_dom[k].append(priv_dom_df.loc[inter])\n",
    "        unpriv_dom[k].append(unpriv_dom_df.loc[inter])\n",
    "\n",
    "priv_dom_mean = [round(np.mean(priv_dom[k])*100) for k in [0,1,2,3]]\n",
    "unpriv_dom_mean = [round(np.mean(unpriv_dom[k])*100) for k in [0,1,2,3]]\n",
    "priv_dom_mean, unpriv_dom_mean "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# percent of times privileged group performed worse than unprivileged group\n",
    "x = np.arange(4)\n",
    "x_lab = [\"Zero\", \"One\", \"Two\", \"Three\"]\n",
    "fig, ax = plt.subplots(figsize=(7, 5))\n",
    "\n",
    "width = 0.35\n",
    "rects1 = ax.bar(x - width/2, priv_dom_mean, label='Privileged', color='steelblue', edgecolor='steelblue', alpha=0.8, width=0.32)\n",
    "rects2 = ax.bar(x + width/2, unpriv_dom_mean, label='Unprivileged', color='orange', edgecolor='orange', alpha=0.8, width=0.32)\n",
    "\n",
    "plt.rcParams.update({'font.size': 14})\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "ax.spines['left'].set_visible(False)\n",
    "ax.spines['bottom'].set_color('#DDDDDD')\n",
    "ax.tick_params(bottom=False, left=False)\n",
    "ax.set_axisbelow(True)\n",
    "ax.yaxis.grid(True, color='#EEEEEE')\n",
    "ax.xaxis.grid(False)\n",
    "ax.bar_label(rects1, padding=3)\n",
    "ax.bar_label(rects2, padding=3)\n",
    "ax.set_ylim([0, 63])\n",
    "ax.set_xticks(x, x_lab)\n",
    "#ax.set_ylabel('% of times group fairness metric <0.01')\n",
    "ax.set_ylabel('% dominance')\n",
    "ax.set_xlabel('Number of Interventions')\n",
    "ax.legend(loc=4, ncol=1)\n",
    "fig.tight_layout()\n",
    "#plt.show()\n",
    "fig.savefig('./figures/priv_unpriv_dom.png', dpi=600)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3>Base Rate</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OP - 3             0.203430\n",
       "OP - 11            0.201515\n",
       "OP - 17            0.212598\n",
       "OP + EOP - 3       0.128713\n",
       "OP + EOP - 11      0.134275\n",
       "                     ...   \n",
       "GSR + CEOP - 11    0.155305\n",
       "GSR + CEOP - 17    0.213572\n",
       "GSR + ROC - 3      0.384903\n",
       "GSR + ROC - 11     0.036168\n",
       "GSR + ROC - 17     0.356460\n",
       "Length: 180, dtype: float64"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adult_df = pandas.read_csv(\"./data/All_Data -adult.csv\", index_col=0)\n",
    "(adult_df[\"Num True Pos - Privileged\"] + adult_df[\"Num False Pos - Privileged\"])/(adult_df[\"Num True Pos - Privileged\"] + adult_df[\"Num True Neg - Privileged\"] + adult_df[\"Num False Pos - Privileged\"] + adult_df[\"Num False Neg - Privileged\"])\n",
    "#adult_df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OP - 3             0.2034\n",
       "OP - 11            0.2015\n",
       "OP - 17            0.2126\n",
       "OP + EOP - 3       0.1287\n",
       "OP + EOP - 11      0.1343\n",
       "                    ...  \n",
       "GSR + CEOP - 11    0.1553\n",
       "GSR + CEOP - 17    0.2136\n",
       "GSR + ROC - 3      0.3849\n",
       "GSR + ROC - 11     0.0362\n",
       "GSR + ROC - 17     0.3565\n",
       "Name: Privileged base Rate, Length: 180, dtype: float64"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adult_df[\"Privileged base Rate\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.concat([adult_df, german_df, bank_df, compas_df])\n",
    "#df = pd.concat([adult_df, german_df, compas_df])\n",
    "df = df[[\"Privileged base Rate\", \"Unprivileged base Rate\", \"base Rate\"]]\n",
    "#df = df.groupby(df.index).mean()\n",
    "\n",
    "#base_rate = {0:{\"priv\": [], \"unpriv\":[]}, 1:{\"priv\": [], \"unpriv\":[]}, 2:{\"priv\": [], \"unpriv\":[]}, 3:{\"priv\": [], \"unpriv\":[]}}\n",
    "base_rate = {0:{\"priv\": [], \"unpriv\":[], \"tot\":[]}, \n",
    "             1:{\"priv\": [], \"unpriv\":[], \"tot\":[]}, \n",
    "             2:{\"priv\": [], \"unpriv\":[], \"tot\":[]}, \n",
    "             3:{\"priv\": [], \"unpriv\":[], \"tot\":[]}}\n",
    "for index, row in df.iterrows():\n",
    "    k = num_interventions(index)\n",
    "    #print(index, k, round(row[\"Privileged base Rate\"],3), round(row[\"Unprivileged base Rate\"],3))\n",
    "    base_rate[k][\"priv\"].append(row[\"Privileged base Rate\"])\n",
    "    base_rate[k][\"unpriv\"].append(row[\"Unprivileged base Rate\"])\n",
    "    base_rate[k][\"tot\"].append(round(row[\"base Rate\"], 4))\n",
    "\n",
    "#base_rate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean\n",
      "priv:  [46, 44, 45, 50] unpriv:  [34, 38, 40, 44] tot:  [41, 42, 43, 48]\n",
      "Std error\n",
      "priv:  [8.53, 3.08, 1.9, 2.03] unpriv:  [6.37, 2.64, 1.76, 2.08] tot:  [7.91, 2.96, 1.85, 2.03]\n"
     ]
    }
   ],
   "source": [
    "priv_base = {}\n",
    "unpriv_base = {}\n",
    "tot_base = {}\n",
    "\n",
    "priv_err = {}\n",
    "unpriv_err = {}\n",
    "tot_err = {}\n",
    "\n",
    "for i in base_rate:\n",
    "    priv_base[i] = round(np.mean(base_rate[i][\"priv\"])*100)\n",
    "    unpriv_base[i] = round(np.mean(base_rate[i][\"unpriv\"])*100)\n",
    "    tot_base[i] = round(np.mean(base_rate[i][\"tot\"])*100)\n",
    "    \n",
    "    priv_err[i] = round(np.std(base_rate[i][\"priv\"])*100/(len(base_rate[i][\"priv\"])**0.5), 2)\n",
    "    unpriv_err[i] = round(np.std(base_rate[i][\"unpriv\"])*100/(len(base_rate[i][\"unpriv\"])**0.5), 2)\n",
    "    tot_err[i] = round(np.std(base_rate[i][\"tot\"])*100/(len(base_rate[i][\"tot\"])**0.5), 2)\n",
    "\n",
    "x = np.arange(4)\n",
    "priv_base = [priv_base[i] for i in x]\n",
    "unpriv_base = [unpriv_base[i] for i in x]\n",
    "tot_base = [tot_base[i] for i in x]\n",
    "x_lab = [\"Zero\", \"One\", \"Two\", \"Three\"]\n",
    "\n",
    "print(\"Mean\")\n",
    "print(\"priv: \",priv_base, \"unpriv: \", unpriv_base, \"tot: \", tot_base)\n",
    "\n",
    "print(\"Std error\")\n",
    "priv_err = [priv_err[i] for i in x]\n",
    "unpriv_err = [unpriv_err[i] for i in x]\n",
    "tot_err = [tot_err[i] for i in x]\n",
    "print(\"priv: \",priv_err, \"unpriv: \", unpriv_err, \"tot: \", tot_err)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "252"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(base_rate[3][\"priv\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x270 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# code: https://www.pythoncharts.com/matplotlib/grouped-bar-charts-matplotlib/\n",
    "width = 0.35  # the width of the bars\n",
    "\n",
    "#fig, ax = plt.subplots(figsize=(7, 5))\n",
    "fig, ax = plt.subplots(figsize=(6, 3.75))\n",
    "plt.rcParams.update({'font.size': 13})\n",
    "\n",
    "c1, c2 = 'royalblue', 'sandybrown'\n",
    "rects1 = ax.bar(x - width/2, priv_base, label='Privileged', yerr=priv_err, color=c1, edgecolor=c1, alpha=0.6, width=0.3, linewidth=2, hatch='.')\n",
    "rects2 = ax.bar(x + width/2, unpriv_base, label='Unprivileged', yerr=unpriv_err, color=c2, edgecolor=c2, alpha=0.6, width=0.3, linewidth=2, hatch='x')\n",
    "\n",
    "# Add some text for labels, title and custom x-axis tick labels, etc.\n",
    "ax.set_ylabel('Mean Base Rate', size=13)\n",
    "ax.set_xlabel('Number of Interventions', size=13)\n",
    "#ax.set_title('Mean Accuracy and F1 Score by number of Interventions')\n",
    "ax.set_xticks(x, x_lab)\n",
    "#ax.set_ylim([0, 58])\n",
    "#ax.legend(bbox_to_anchor=(0.5, 1.13), loc='upper center', ncol=2)\n",
    "ax.legend(loc='lower right', ncol=1)\n",
    "#ax.legend(bbox_to_anchor=(1.05, 1), ncol=1)\n",
    "\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "ax.spines['left'].set_visible(False)\n",
    "ax.spines['bottom'].set_color('#DDDDDD')\n",
    "ax.tick_params(bottom=False, left=False)\n",
    "ax.set_axisbelow(True)\n",
    "ax.yaxis.grid(True, color='#EEEEEE')\n",
    "ax.xaxis.grid(False)\n",
    "\n",
    "ax.bar_label(rects1, padding=3)\n",
    "ax.bar_label(rects2, padding=3)\n",
    "\n",
    "fig.tight_layout()\n",
    "plt.show()\n",
    "#fig.savefig('./figures/base_rate.png', dpi=600)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3>Privileged group gets Discriminated</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "adult_df = pd.read_csv(\"./data/adult.csv\", index_col=0)\n",
    "compas_df = pd.read_csv(\"./data/compas.csv\", index_col=0)\n",
    "bank_df = pd.read_csv(\"./data/bank.csv\", index_col=0)\n",
    "german_df = pd.read_csv(\"./data/german.csv\", index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "def grouped_bar_chart(xlabels, y1, y2, title):\n",
    "    # code: https://www.pythoncharts.com/matplotlib/grouped-bar-charts-matplotlib/\n",
    "    width = 0.35  # the width of the bars\n",
    "    x = np.arange(len(xlabels))\n",
    "    fig, ax = plt.subplots(figsize=(18, 8))\n",
    "    plt.rcParams.update({'font.size': 15})\n",
    "    #c1, c2 = 'royalblue', 'darksalmon'\n",
    "    c1, c2 = 'royalblue', 'sandybrown'\n",
    "    rects1 = ax.bar(x - width/2, y1, label='Privileged', color=c1, edgecolor=c1, alpha=0.6, width=0.3, linewidth=2, hatch='.')\n",
    "    rects2 = ax.bar(x + width/2, y2, label='Unprivileged', color=c2, edgecolor=c2, alpha=0.6, width=0.3, linewidth=2, hatch='x')\n",
    "\n",
    "    # Add some text for labels, title and custom x-axis tick labels, etc.\n",
    "    ax.set_ylabel('difference in FNR compared to baseline', size=17)\n",
    "    #ax.set_xlabel('Number of Interventions')\n",
    "    #ax.set_title('Change in False Positive Rate across different Interventions for COMPAS dataset', size=16)\n",
    "    #ax.set_title(title, size=16)\n",
    "    ax.set_xticks(x, xlabels, rotation = 90, size=15)\n",
    "    ax.set_ylim([-29, 29])\n",
    "    ax.legend()\n",
    "\n",
    "    ax.spines['top'].set_visible(False)\n",
    "    ax.spines['right'].set_visible(False)\n",
    "    ax.spines['left'].set_visible(False)\n",
    "    ax.spines['bottom'].set_color('#DDDDDD')\n",
    "    #ax.tick_params(bottom=False, left=False)\n",
    "    ax.set_axisbelow(True)\n",
    "    ax.yaxis.grid(True, color='#EEEEEE')\n",
    "    ax.xaxis.grid(False)\n",
    "\n",
    "    ax.bar_label(rects1, padding=3)\n",
    "    ax.bar_label(rects2, padding=3)\n",
    "\n",
    "    fig.tight_layout()\n",
    "    #plt.show()\n",
    "    #fig.savefig('./figures/base_rate.png', dpi=600)\n",
    "    return fig"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-8.813 10.64 1.827\n",
      "-8.813 10.61 1.797\n",
      "-8.813 8.7 -0.117\n",
      "-8.813 10.64 1.827\n",
      "-8.813 -4.23 -13.047\n",
      "-8.813 17.66 8.847\n",
      "-8.813 10.27 1.453\n",
      "-8.813 3.48 -5.337\n",
      "-8.813 17.04 8.23\n",
      "-8.813 15.66 6.843\n",
      "-8.813 8.36 -0.457\n",
      "-8.813 9.17 0.353\n",
      "-8.813 8.73 -0.08\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Let's go with Compas dataset - False Positive Rate\n",
    "# For compas dataset, positive means - No recidivate (label:0)\n",
    "# so false positive means - Actually committed crime but labeled as innocent\n",
    "\n",
    "#for data in [adult_df, compas_df, german_df, bank_df]:\n",
    "for data in [adult_df]:\n",
    "    #cols = [\"False Positive Rate\", \"False Negative Rate\", \"False Discovery Rate\", \"False Omission Rate\"]\n",
    "    cols = [\"False Negative Rate\"]\n",
    "    for metric in cols:\n",
    "        b1 = data[metric + \" - Privileged\"][\"Logistic Regression\"]\n",
    "        b2 = data[metric + \" - Unprivileged\"][\"Logistic Regression\"]\n",
    "        #b1 = data['Privileged base Rate'][\"Logistic Regression\"]\n",
    "        #b2 = data['Unprivileged base Rate'][\"Logistic Regression\"]\n",
    "        m_ori = b1 - b2    # original metric value FPR Diff\n",
    "        res = {\"priv\":[], \"unpriv\": []}\n",
    "        row_names = []\n",
    "        for index, row in data.iterrows():\n",
    "            v1 = row[metric + \" - Privileged\"]\n",
    "            v2 = row[metric + \" - Unprivileged\"]\n",
    "            #v1 = row['Privileged base Rate']\n",
    "            #v2 = row['Unprivileged base Rate']\n",
    "            \n",
    "            if index in [\"OP + PR + CEOP\", \"OP + PR\"]:\n",
    "                continue\n",
    "            \n",
    "            m_new = v1 - v2    # new value for FPR Diff\n",
    "            d1, d2 = v1-b1, v2-b2  # delta change in value\n",
    "            #if abs(d1)>0.05 and abs(d2)>0.05 and d1*d2<0:\n",
    "            if abs(m_new)-abs(m_ori)<0.05 and 0.03<abs(d1)<0.26 and 0.03<abs(d2)<0.32:\n",
    "            #if abs(m_new)<abs(m_ori) and d1>0.01 and d2<0.01:\n",
    "                print(round(m_ori*100,3), round((d1-d2)*100,2),round(m_new*100,3))\n",
    "                res[\"priv\"].append(round(d1,2)*100)\n",
    "                res[\"unpriv\"].append(round(d2,2)*100)\n",
    "                row_names.append(index)\n",
    "\n",
    "        fig = grouped_bar_chart(row_names, res[\"priv\"], res[\"unpriv\"], metric)\n",
    "        #fig.savefig('./figures/fpr_compas.png', dpi=600)\n",
    "        fig.savefig('./figures/fnr_adult.png', dpi=600)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:root:Missing Data: 5 rows removed from CompasDataset.\n"
     ]
    },
    {
     "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>two_year_recid</th>\n",
       "      <th>0.0</th>\n",
       "      <th>1.0</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>race</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.0</th>\n",
       "      <td>2080</td>\n",
       "      <td>1987</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1.0</th>\n",
       "      <td>1278</td>\n",
       "      <td>822</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "two_year_recid   0.0   1.0\n",
       "race                      \n",
       "0.0             2080  1987\n",
       "1.0             1278   822"
      ]
     },
     "execution_count": 138,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#pd.crosstab(compas_df[])\n",
    "from aif360.datasets import CompasDataset\n",
    "c = CompasDataset().convert_to_dataframe()[0]\n",
    "pd.crosstab(c[\"race\"], c[\"two_year_recid\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3>Error rate gets worse</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "adult_df = pd.read_csv(\"./data/adult.csv\", index_col=0)\n",
    "compas_df = pd.read_csv(\"./data/compas.csv\", index_col=0)\n",
    "bank_df = pd.read_csv(\"./data/bank.csv\", index_col=0)\n",
    "german_df = pd.read_csv(\"./data/german.csv\", index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_plot(y1, y2, metric):\n",
    "    xlabels = [\"One\", \"Two\", \"Three\"]\n",
    "\n",
    "    width = 0.35  # the width of the bars\n",
    "    x = np.arange(len(xlabels))\n",
    "    fig, ax = plt.subplots(figsize=(5, 3.75))\n",
    "    plt.rcParams.update({'font.size': 13})\n",
    "    c1, c2 = 'royalblue', 'sandybrown'\n",
    "    rects1 = ax.bar(x - width/2, y1, label='Privileged', color=c1, edgecolor=c1, alpha=0.6, width=0.3, linewidth=2, hatch='.')\n",
    "    rects2 = ax.bar(x + width/2, y2, label='Unprivileged', color=c2, edgecolor=c2, alpha=0.6, width=0.3, linewidth=2, hatch='x')\n",
    "\n",
    "    # Add some text for labels, title and custom x-axis tick labels, etc.\n",
    "    ylabel = ''\n",
    "    if metric==\"F1 Score\":\n",
    "        ylabel = '% times f1 score improves'\n",
    "    elif metric==\"False Positive Rate\":\n",
    "        ylabel = '% times FPR increases'\n",
    "    elif metric==\"False Negative Rate\":\n",
    "        ylabel = '% times FNR increases'\n",
    "    ax.set_ylabel(ylabel, size=13)\n",
    "    ax.set_xlabel('Number of Interventions')\n",
    "    #ax.set_title('Change in False Positive Rate across different Interventions for COMPAS dataset', size=16)\n",
    "    #ax.set_title(title, size=16)\n",
    "    ax.set_xticks(x, xlabels, rotation = 0, size=13)\n",
    "    #ax.set_ylim([30, 70])\n",
    "    if metric==\"False Positive Rate\":\n",
    "        #ax.set_ylim([0, 30])\n",
    "        ax.legend(loc='lower right')\n",
    "        #ax.legend(loc='upper right', bbox_to_anchor=(1, 1))\n",
    "        #ax.legend(loc='lower right', bbox_to_anchor=(1, 0))\n",
    "\n",
    "    ax.spines['top'].set_visible(False)\n",
    "    ax.spines['right'].set_visible(False)\n",
    "    ax.spines['left'].set_visible(False)\n",
    "    ax.spines['bottom'].set_color('#DDDDDD')\n",
    "    #ax.tick_params(bottom=False, left=False)\n",
    "    ax.set_axisbelow(True)\n",
    "    ax.yaxis.grid(True, color='#EEEEEE')\n",
    "    ax.xaxis.grid(False)\n",
    "\n",
    "    ax.bar_label(rects1, padding=3)\n",
    "    ax.bar_label(rects2, padding=3)\n",
    "\n",
    "    fig.tight_layout()\n",
    "    fig.savefig('./figures/'+metric+'.png', dpi=600)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Metric:  False Positive Rate\n",
      "{'priv': [13, 43, 41], 'unpriv': [20, 65, 60]}\n",
      "[37, 44, 49] [57, 67, 71]\n",
      "Metric:  False Negative Rate\n",
      "{'priv': [20, 68, 58], 'unpriv': [16, 55, 44]}\n",
      "[57, 70, 69] [46, 57, 52]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x270 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x270 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#cols = [\"False Positive Rate\", \"False Negative Rate\", \"False Discovery Rate\", \"False Omission Rate\", \"Accuracy\"]\n",
    "cols = [\"False Positive Rate\", \"False Negative Rate\"]\n",
    "    \n",
    "for metric in cols:  \n",
    "    print(\"Metric: \", metric)\n",
    "    res = {\"priv\":[0, 0, 0], \"unpriv\": [0, 0, 0]}\n",
    "    tot_cases = [0, 0, 0]\n",
    "\n",
    "    for data in [adult_df, compas_df, german_df, bank_df]:\n",
    "    #for data in [adult_df]:\n",
    "\n",
    "        base_p = data[metric + \" - Privileged\"][\"Logistic Regression\"]\n",
    "        base_u = data[metric + \" - Unprivileged\"][\"Logistic Regression\"]\n",
    "        \n",
    "    \n",
    "        for index, row in data.iterrows():\n",
    "            if index == \"Logistic Regression\":\n",
    "                continue\n",
    "                \n",
    "            new_p = row[metric + \" - Privileged\"]\n",
    "            new_u = row[metric + \" - Unprivileged\"]\n",
    "            \n",
    "            n = num_interventions(index)\n",
    "            tot_cases[n-1] += 1\n",
    "            \n",
    "            if metric not in [\"F1 Score\", \"Accuracy\"]:     \n",
    "                    \n",
    "                if abs(new_p)>abs(base_p):\n",
    "                    res[\"priv\"][n-1] += 1\n",
    "\n",
    "                if abs(new_u)>abs(base_u):\n",
    "                    res[\"unpriv\"][n-1] += 1\n",
    "            else:\n",
    "                \n",
    "                if abs(new_p)>=abs(base_p):\n",
    "                    res[\"priv\"][n-1] += 1\n",
    "\n",
    "                if abs(new_u)>=abs(base_u):\n",
    "                    res[\"unpriv\"][n-1] += 1\n",
    "        \n",
    "    print(res)\n",
    "    priv_per = [round(res[\"priv\"][i]/tot_cases[i]*100) for i in [0,1,2]]\n",
    "    unpriv_per = [round(res[\"unpriv\"][i]/tot_cases[i]*100) for i in [0,1,2]]\n",
    "    print(priv_per, unpriv_per)\n",
    "    make_plot(priv_per, unpriv_per, metric)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h3>F1 score and Accuracy by group</h3>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(660, 30)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#adult_df = pd.read_csv(\"./data/adult.csv\", index_col=0)\n",
    "#compas_df = pd.read_csv(\"./data/compas.csv\", index_col=0)\n",
    "#bank_df = pd.read_csv(\"./data/bank.csv\", index_col=0)\n",
    "#german_df = pd.read_csv(\"./data/german.csv\", index_col=0)\n",
    "\n",
    "adult_df = pandas.read_csv(\"./data/All_Data -adult.csv\", index_col=0)\n",
    "compas_df = pandas.read_csv(\"./data/All_Data -compas.csv\", index_col=0)\n",
    "bank_df = pandas.read_csv(\"./data/All_Data -bank.csv\", index_col=0)\n",
    "german_df = pandas.read_csv(\"./data/All_Data -german.csv\", index_col=0)\n",
    "\n",
    "all_data = pd.concat([adult_df, compas_df, bank_df, german_df])\n",
    "#all_data = pd.concat([adult_df])\n",
    "all_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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></th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>F1 Score</th>\n",
       "      <th>Theil Index</th>\n",
       "      <th>False Positive Rate - Unprivileged</th>\n",
       "      <th>False Positive Rate - Privileged</th>\n",
       "      <th>False Negative Rate - Unprivileged</th>\n",
       "      <th>False Negative Rate - Privileged</th>\n",
       "      <th>Accuracy - Unprivileged</th>\n",
       "      <th>Accuracy - Privileged</th>\n",
       "      <th>False Discovery Rate - Unprivileged</th>\n",
       "      <th>...</th>\n",
       "      <th>Num True Pos - Unprivileged</th>\n",
       "      <th>Num True Neg - Unprivileged</th>\n",
       "      <th>Num False Pos - Unprivileged</th>\n",
       "      <th>Num False Neg - Unprivileged</th>\n",
       "      <th>F1 Score - Privileged</th>\n",
       "      <th>F1 Score - Unprivileged</th>\n",
       "      <th>Privileged base Rate</th>\n",
       "      <th>Unprivileged base Rate</th>\n",
       "      <th>Consistency</th>\n",
       "      <th>num_inter</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>OP - 3</th>\n",
       "      <td>0.7878</td>\n",
       "      <td>0.4844</td>\n",
       "      <td>0.1840</td>\n",
       "      <td>0.0628</td>\n",
       "      <td>0.1015</td>\n",
       "      <td>0.7212</td>\n",
       "      <td>0.5744</td>\n",
       "      <td>0.8643</td>\n",
       "      <td>0.7498</td>\n",
       "      <td>0.6437</td>\n",
       "      <td>...</td>\n",
       "      <td>150.0</td>\n",
       "      <td>4047.0</td>\n",
       "      <td>271.0</td>\n",
       "      <td>388.0</td>\n",
       "      <td>0.5169</td>\n",
       "      <td>0.3128</td>\n",
       "      <td>0.2034</td>\n",
       "      <td>0.0867</td>\n",
       "      <td>0.9842</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP - 11</th>\n",
       "      <td>0.7898</td>\n",
       "      <td>0.4764</td>\n",
       "      <td>0.1788</td>\n",
       "      <td>0.0680</td>\n",
       "      <td>0.1031</td>\n",
       "      <td>0.7371</td>\n",
       "      <td>0.5743</td>\n",
       "      <td>0.8632</td>\n",
       "      <td>0.7531</td>\n",
       "      <td>0.6930</td>\n",
       "      <td>...</td>\n",
       "      <td>132.0</td>\n",
       "      <td>4082.0</td>\n",
       "      <td>298.0</td>\n",
       "      <td>370.0</td>\n",
       "      <td>0.5127</td>\n",
       "      <td>0.2833</td>\n",
       "      <td>0.2015</td>\n",
       "      <td>0.0881</td>\n",
       "      <td>0.9687</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP - 17</th>\n",
       "      <td>0.7934</td>\n",
       "      <td>0.4929</td>\n",
       "      <td>0.1707</td>\n",
       "      <td>0.0703</td>\n",
       "      <td>0.1102</td>\n",
       "      <td>0.7135</td>\n",
       "      <td>0.5456</td>\n",
       "      <td>0.8590</td>\n",
       "      <td>0.7603</td>\n",
       "      <td>0.6652</td>\n",
       "      <td>...</td>\n",
       "      <td>155.0</td>\n",
       "      <td>4072.0</td>\n",
       "      <td>308.0</td>\n",
       "      <td>386.0</td>\n",
       "      <td>0.5301</td>\n",
       "      <td>0.3088</td>\n",
       "      <td>0.2126</td>\n",
       "      <td>0.0941</td>\n",
       "      <td>0.9834</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP + EOP - 3</th>\n",
       "      <td>0.7727</td>\n",
       "      <td>0.3717</td>\n",
       "      <td>0.2177</td>\n",
       "      <td>0.0628</td>\n",
       "      <td>0.0634</td>\n",
       "      <td>0.7212</td>\n",
       "      <td>0.7289</td>\n",
       "      <td>0.8643</td>\n",
       "      <td>0.7274</td>\n",
       "      <td>0.6437</td>\n",
       "      <td>...</td>\n",
       "      <td>150.0</td>\n",
       "      <td>4047.0</td>\n",
       "      <td>271.0</td>\n",
       "      <td>388.0</td>\n",
       "      <td>0.3847</td>\n",
       "      <td>0.3128</td>\n",
       "      <td>0.1287</td>\n",
       "      <td>0.0867</td>\n",
       "      <td>0.9211</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP + EOP - 11</th>\n",
       "      <td>0.7737</td>\n",
       "      <td>0.3698</td>\n",
       "      <td>0.2099</td>\n",
       "      <td>0.0760</td>\n",
       "      <td>0.0697</td>\n",
       "      <td>0.7331</td>\n",
       "      <td>0.7186</td>\n",
       "      <td>0.8564</td>\n",
       "      <td>0.7324</td>\n",
       "      <td>0.7131</td>\n",
       "      <td>...</td>\n",
       "      <td>134.0</td>\n",
       "      <td>4047.0</td>\n",
       "      <td>333.0</td>\n",
       "      <td>368.0</td>\n",
       "      <td>0.3909</td>\n",
       "      <td>0.2766</td>\n",
       "      <td>0.1343</td>\n",
       "      <td>0.0957</td>\n",
       "      <td>0.9175</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "               Accuracy  F1 Score  Theil Index  \\\n",
       "OP - 3           0.7878    0.4844       0.1840   \n",
       "OP - 11          0.7898    0.4764       0.1788   \n",
       "OP - 17          0.7934    0.4929       0.1707   \n",
       "OP + EOP - 3     0.7727    0.3717       0.2177   \n",
       "OP + EOP - 11    0.7737    0.3698       0.2099   \n",
       "\n",
       "               False Positive Rate - Unprivileged  \\\n",
       "OP - 3                                     0.0628   \n",
       "OP - 11                                    0.0680   \n",
       "OP - 17                                    0.0703   \n",
       "OP + EOP - 3                               0.0628   \n",
       "OP + EOP - 11                              0.0760   \n",
       "\n",
       "               False Positive Rate - Privileged  \\\n",
       "OP - 3                                   0.1015   \n",
       "OP - 11                                  0.1031   \n",
       "OP - 17                                  0.1102   \n",
       "OP + EOP - 3                             0.0634   \n",
       "OP + EOP - 11                            0.0697   \n",
       "\n",
       "               False Negative Rate - Unprivileged  \\\n",
       "OP - 3                                     0.7212   \n",
       "OP - 11                                    0.7371   \n",
       "OP - 17                                    0.7135   \n",
       "OP + EOP - 3                               0.7212   \n",
       "OP + EOP - 11                              0.7331   \n",
       "\n",
       "               False Negative Rate - Privileged  Accuracy - Unprivileged  \\\n",
       "OP - 3                                   0.5744                   0.8643   \n",
       "OP - 11                                  0.5743                   0.8632   \n",
       "OP - 17                                  0.5456                   0.8590   \n",
       "OP + EOP - 3                             0.7289                   0.8643   \n",
       "OP + EOP - 11                            0.7186                   0.8564   \n",
       "\n",
       "               Accuracy - Privileged  False Discovery Rate - Unprivileged  \\\n",
       "OP - 3                        0.7498                               0.6437   \n",
       "OP - 11                       0.7531                               0.6930   \n",
       "OP - 17                       0.7603                               0.6652   \n",
       "OP + EOP - 3                  0.7274                               0.6437   \n",
       "OP + EOP - 11                 0.7324                               0.7131   \n",
       "\n",
       "               ...  Num True Pos - Unprivileged  Num True Neg - Unprivileged  \\\n",
       "OP - 3         ...                        150.0                       4047.0   \n",
       "OP - 11        ...                        132.0                       4082.0   \n",
       "OP - 17        ...                        155.0                       4072.0   \n",
       "OP + EOP - 3   ...                        150.0                       4047.0   \n",
       "OP + EOP - 11  ...                        134.0                       4047.0   \n",
       "\n",
       "               Num False Pos - Unprivileged  Num False Neg - Unprivileged  \\\n",
       "OP - 3                                271.0                         388.0   \n",
       "OP - 11                               298.0                         370.0   \n",
       "OP - 17                               308.0                         386.0   \n",
       "OP + EOP - 3                          271.0                         388.0   \n",
       "OP + EOP - 11                         333.0                         368.0   \n",
       "\n",
       "               F1 Score - Privileged  F1 Score - Unprivileged  \\\n",
       "OP - 3                        0.5169                   0.3128   \n",
       "OP - 11                       0.5127                   0.2833   \n",
       "OP - 17                       0.5301                   0.3088   \n",
       "OP + EOP - 3                  0.3847                   0.3128   \n",
       "OP + EOP - 11                 0.3909                   0.2766   \n",
       "\n",
       "               Privileged base Rate  Unprivileged base Rate  Consistency  \\\n",
       "OP - 3                       0.2034                  0.0867       0.9842   \n",
       "OP - 11                      0.2015                  0.0881       0.9687   \n",
       "OP - 17                      0.2126                  0.0941       0.9834   \n",
       "OP + EOP - 3                 0.1287                  0.0867       0.9211   \n",
       "OP + EOP - 11                0.1343                  0.0957       0.9175   \n",
       "\n",
       "               num_inter  \n",
       "OP - 3                 1  \n",
       "OP - 11                1  \n",
       "OP - 17                1  \n",
       "OP + EOP - 3           2  \n",
       "OP + EOP - 11          2  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_data[\"num_inter\"] = [num_interventions(n) for n in all_data.index]\n",
    "all_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[68, 62, 58, 59] [3.38, 1.76, 1.26, 1.31]\n",
      "[79, 75, 72, 70] [2.66, 1.26, 0.83, 0.82]\n"
     ]
    }
   ],
   "source": [
    "f1_mean = all_data.groupby(\"num_inter\").mean()[\"F1 Score\"].tolist()\n",
    "f1_mean = [round(x*100) for x in f1_mean]\n",
    "f1_err = all_data.groupby(\"num_inter\").std()[\"F1 Score\"]/(all_data.groupby(\"num_inter\").size()**0.5).tolist()\n",
    "f1_err = [round(x*100, 2) for x in f1_err]\n",
    "print(f1_mean, f1_err)\n",
    "\n",
    "acc_mean = all_data.groupby(\"num_inter\").mean()[\"Accuracy\"].tolist()\n",
    "acc_mean = [round(x*100) for x in acc_mean]\n",
    "acc_err = all_data.groupby(\"num_inter\").std()[\"Accuracy\"]/(all_data.groupby(\"num_inter\").size()**0.5).tolist()\n",
    "acc_err = [round(x*100, 2) for x in acc_err]\n",
    "print(acc_mean, acc_err)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "width = 0.35  # the width of the bars\n",
    "x = x = np.arange(4)\n",
    "labels = [\"Zero\", \"One\", \"Two\", \"Three\"]\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(7, 5))\n",
    "plt.rcParams.update({'font.size': 15})\n",
    "rects1 = ax.bar(x - width/2, acc_mean, label='Accuracy', yerr=acc_err, color='steelblue', edgecolor='steelblue', alpha=0.6, width=0.32, hatch='x')\n",
    "rects2 = ax.bar(x + width/2, f1_mean, label='F1 Score', yerr=f1_err, color='lightgrey', edgecolor='darkgrey', alpha=0.6, width=0.32, hatch='.')\n",
    "\n",
    "# Add some text for labels, title and custom x-axis tick labels, etc.\n",
    "#ax.set_ylabel('Metric value', size=15)\n",
    "ax.set_ylabel('Percent value', size=15)\n",
    "ax.set_xlabel('Number of Interventions', size=15)\n",
    "#ax.set_title('Mean Accuracy and F1 Score by number of Interventions')\n",
    "ax.set_xticks(x, labels, size=15)\n",
    "ax.set_ylim([0, 88])\n",
    "ax.legend(bbox_to_anchor=(0.5, 1.13), loc='upper center', ncol=2)\n",
    "\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "ax.spines['left'].set_visible(False)\n",
    "ax.spines['bottom'].set_color('#DDDDDD')\n",
    "ax.tick_params(bottom=False, left=False)\n",
    "ax.set_axisbelow(True)\n",
    "ax.yaxis.grid(True, color='#EEEEEE')\n",
    "ax.xaxis.grid(False)\n",
    "\n",
    "ax.bar_label(rects1, padding=3)\n",
    "ax.bar_label(rects2, padding=3)\n",
    "\n",
    "fig.tight_layout()\n",
    "#plt.show()\n",
    "fig.savefig('./figures/utility_by_intervention.png', dpi=600)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [],
   "source": [
    "f1_p = all_data.groupby(\"num_inter\").mean()[\"F1 Score - Privileged\"].tolist()\n",
    "f1_u = all_data.groupby(\"num_inter\").mean()[\"F1 Score - Unprivileged\"].tolist()\n",
    "\n",
    "f1_p_err = all_data.groupby(\"num_inter\").std()[\"F1 Score - Privileged\"]/(all_data.groupby(\"num_inter\").size()**0.5).tolist()\n",
    "f1_u_err = all_data.groupby(\"num_inter\").std()[\"F1 Score - Unprivileged\"]/(all_data.groupby(\"num_inter\").size()**0.5).tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[69, 64, 60, 61] [63, 56, 51, 49]\n",
      "[3.73, 1.9, 1.34, 1.39] [1.78, 1.6, 1.22, 1.38]\n"
     ]
    }
   ],
   "source": [
    "f1_p = [round(x*100) for x in f1_p]\n",
    "f1_u = [round(x*100) for x in f1_u]\n",
    "\n",
    "f1_p_err = [round(x*100,2) for x in f1_p_err]\n",
    "f1_u_err = [round(x*100,2) for x in f1_u_err]\n",
    "\n",
    "print(f1_p, f1_u)\n",
    "print(f1_p_err, f1_u_err)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ttest_indResult(statistic=1.5353213658105023, pvalue=0.1389624579303136)\n",
      "Ttest_indResult(statistic=3.1071462173795683, pvalue=0.0021528429492939545)\n",
      "Ttest_indResult(statistic=5.055043585935102, pvalue=5.777544913666047e-07)\n",
      "Ttest_indResult(statistic=6.336682688313132, pvalue=5.227918191907225e-10)\n",
      "Accuracy\n",
      "Ttest_indResult(statistic=0.44464271549725687, pvalue=0.6609215621824925)\n",
      "Ttest_indResult(statistic=1.612842601438119, pvalue=0.10829487772987226)\n",
      "Ttest_indResult(statistic=2.1796538048361347, pvalue=0.029684619366500572)\n",
      "Ttest_indResult(statistic=2.464919859085577, pvalue=0.014038073980346878)\n"
     ]
    }
   ],
   "source": [
    "for num_interven in [0,1,2,3]:\n",
    "    p = all_data[all_data[\"num_inter\"]==num_interven][\"F1 Score - Privileged\"]\n",
    "    u = all_data[all_data[\"num_inter\"]==num_interven][\"F1 Score - Unprivileged\"]\n",
    "    print(ttest_ind(p, u))\n",
    "    \n",
    "print(\"Accuracy\")\n",
    "for num_interven in [0,1,2,3]:\n",
    "    p = all_data[all_data[\"num_inter\"]==num_interven][\"Accuracy - Privileged\"]\n",
    "    u = all_data[all_data[\"num_inter\"]==num_interven][\"Accuracy - Unprivileged\"]\n",
    "    print(ttest_ind(p, u))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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></th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Accuracy - Unprivileged</th>\n",
       "      <th>Accuracy - Privileged</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>OP</th>\n",
       "      <td>0.790333</td>\n",
       "      <td>0.862167</td>\n",
       "      <td>0.754400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP + EOP</th>\n",
       "      <td>0.774533</td>\n",
       "      <td>0.859400</td>\n",
       "      <td>0.732133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP + CEOP</th>\n",
       "      <td>0.790333</td>\n",
       "      <td>0.862167</td>\n",
       "      <td>0.754400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP + ROC</th>\n",
       "      <td>0.684867</td>\n",
       "      <td>0.645900</td>\n",
       "      <td>0.704433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OP + GFC</th>\n",
       "      <td>0.786900</td>\n",
       "      <td>0.864600</td>\n",
       "      <td>0.748033</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EGR + ROC</th>\n",
       "      <td>0.741133</td>\n",
       "      <td>0.591000</td>\n",
       "      <td>0.778033</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GSR</th>\n",
       "      <td>0.742233</td>\n",
       "      <td>0.616167</td>\n",
       "      <td>0.773933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GSR + EOP</th>\n",
       "      <td>0.713367</td>\n",
       "      <td>0.616167</td>\n",
       "      <td>0.737767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GSR + CEOP</th>\n",
       "      <td>0.725567</td>\n",
       "      <td>0.616167</td>\n",
       "      <td>0.753033</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GSR + ROC</th>\n",
       "      <td>0.615567</td>\n",
       "      <td>0.692200</td>\n",
       "      <td>0.595267</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>220 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Accuracy  Accuracy - Unprivileged  Accuracy - Privileged\n",
       "OP          0.790333                 0.862167               0.754400\n",
       "OP + EOP    0.774533                 0.859400               0.732133\n",
       "OP + CEOP   0.790333                 0.862167               0.754400\n",
       "OP + ROC    0.684867                 0.645900               0.704433\n",
       "OP + GFC    0.786900                 0.864600               0.748033\n",
       "...              ...                      ...                    ...\n",
       "EGR + ROC   0.741133                 0.591000               0.778033\n",
       "GSR         0.742233                 0.616167               0.773933\n",
       "GSR + EOP   0.713367                 0.616167               0.737767\n",
       "GSR + CEOP  0.725567                 0.616167               0.753033\n",
       "GSR + ROC   0.615567                 0.692200               0.595267\n",
       "\n",
       "[220 rows x 3 columns]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_data[[\"Accuracy\", \"Accuracy - Unprivileged\", \"Accuracy - Privileged\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7925 0.7927 0.7927\n",
      "0.6388 0.6389 0.6389\n",
      "0.8488 0.8488 0.8496\n"
     ]
    }
   ],
   "source": [
    "# checking if net accuracy is always in between of the two groups\n",
    "for index, row in all_data.iterrows():\n",
    "    max_acc = max(row[\"Accuracy - Unprivileged\"], row[\"Accuracy - Privileged\"])\n",
    "    min_acc = min(row[\"Accuracy - Unprivileged\"], row[\"Accuracy - Privileged\"])\n",
    "    if not (min_acc<row[\"Accuracy\"] and row[\"Accuracy\"]<max_acc):\n",
    "        print(min_acc, row[\"Accuracy\"], max_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "num_inter\n",
       "0    0.792525\n",
       "1    0.754819\n",
       "2    0.724473\n",
       "3    0.700744\n",
       "Name: Accuracy, dtype: float64"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_data.groupby(\"num_inter\").mean()[\"Accuracy\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "num_inter\n",
       "0    0.092168\n",
       "1    0.128837\n",
       "2    0.140748\n",
       "3    0.129901\n",
       "Name: Accuracy, dtype: float64"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_data.groupby(\"num_inter\").std()[\"Accuracy\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([0.7874083333333334,\n",
       "  0.7545428571428572,\n",
       "  0.7239841924398625,\n",
       "  0.7045785714285714],\n",
       " [0.7682833333333333, 0.7238790476190476, 0.6970676975945017, 0.673325])"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acc_p = all_data.groupby(\"num_inter\").mean()[\"Accuracy - Privileged\"].tolist()\n",
    "acc_u = all_data.groupby(\"num_inter\").mean()[\"Accuracy - Unprivileged\"].tolist()\n",
    "\n",
    "acc_p, acc_u"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[79, 75, 72, 70] [77, 72, 70, 67]\n",
      "[2.62, 1.22, 0.83, 0.8] [3.41, 1.45, 0.91, 0.98]\n"
     ]
    }
   ],
   "source": [
    "acc_p = all_data.groupby(\"num_inter\").mean()[\"Accuracy - Privileged\"].tolist()\n",
    "acc_u = all_data.groupby(\"num_inter\").mean()[\"Accuracy - Unprivileged\"].tolist()\n",
    "\n",
    "acc_p = [round(x*100) for x in acc_p]\n",
    "acc_u = [round(x*100) for x in acc_u]\n",
    "\n",
    "acc_err_p = all_data.groupby(\"num_inter\").std()[\"Accuracy - Privileged\"]/(all_data.groupby(\"num_inter\").size()**0.5).tolist()\n",
    "acc_err_u = all_data.groupby(\"num_inter\").std()[\"Accuracy - Unprivileged\"]/(all_data.groupby(\"num_inter\").size()**0.5).tolist()\n",
    "\n",
    "acc_err_p = [round(x*100,2) for x in acc_err_p]\n",
    "acc_err_u = [round(x*100,2) for x in acc_err_u]\n",
    "\n",
    "print(acc_p, acc_u)\n",
    "print(acc_err_p, acc_err_u)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "xlabels = [\"Zero\", \"One\", \"Two\", \"Three\"]\n",
    "width = 0.35  # the width of the bars\n",
    "x = np.arange(len(xlabels))\n",
    "fig, ax = plt.subplots(figsize=(7, 5))\n",
    "plt.rcParams.update({'font.size': 15})\n",
    "#c1, c2 = 'royalblue', 'darksalmon'\n",
    "c1, c2 = 'royalblue', 'sandybrown'\n",
    "rects1 = ax.bar(x - width/2, f1_p, label='Privileged', yerr=f1_p_err, color=c1, edgecolor=c1, alpha=0.6, width=0.3, linewidth=2, hatch='.')\n",
    "rects2 = ax.bar(x + width/2, f1_u, label='Unprivileged', yerr=f1_u_err, color=c2, edgecolor=c2, alpha=0.6, width=0.3, linewidth=2, hatch='x')\n",
    "\n",
    "# Add some text for labels, title and custom x-axis tick labels, etc.\n",
    "ax.set_ylabel('Mean F1 score', size=15)\n",
    "ax.set_xlabel('Number of Interventions')\n",
    "    #ax.set_title('Change in False Positive Rate across different Interventions for COMPAS dataset', size=16)\n",
    "    #ax.set_title(title, size=16)\n",
    "ax.set_xticks(x, xlabels, rotation = 0, size=15)\n",
    "ax.set_ylim([0, 78])\n",
    "ax.legend(bbox_to_anchor=(0.5, 1.13), loc='upper center', ncol=2)\n",
    "\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "ax.spines['left'].set_visible(False)\n",
    "ax.spines['bottom'].set_color('#DDDDDD')\n",
    "#ax.tick_params(bottom=False, left=False)\n",
    "ax.set_axisbelow(True)\n",
    "ax.yaxis.grid(True, color='#EEEEEE')\n",
    "ax.xaxis.grid(False)\n",
    "\n",
    "ax.bar_label(rects1, padding=3)\n",
    "ax.bar_label(rects2, padding=3)\n",
    "\n",
    "fig.tight_layout()\n",
    "#plt.show()\n",
    "fig.savefig('./figures/f1_score.png', dpi=600)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "xlabels = [\"Zero\", \"One\", \"Two\", \"Three\"]\n",
    "width = 0.35  # the width of the bars\n",
    "x = np.arange(len(xlabels))\n",
    "fig, ax = plt.subplots(figsize=(7, 5))\n",
    "plt.rcParams.update({'font.size': 15})\n",
    "#c1, c2 = 'royalblue', 'darksalmon'\n",
    "c1, c2 = 'royalblue', 'sandybrown'\n",
    "rects1 = ax.bar(x - width/2, acc_p, label='Privileged', yerr=acc_err_p, color=c1, edgecolor=c1, alpha=0.6, width=0.3, linewidth=2, hatch='.')\n",
    "rects2 = ax.bar(x + width/2, acc_u, label='Unprivileged', yerr=acc_err_u, color=c2, edgecolor=c2, alpha=0.6, width=0.3, linewidth=2, hatch='x')\n",
    "\n",
    "# Add some text for labels, title and custom x-axis tick labels, etc.\n",
    "ax.set_ylabel('Mean Accuracy score', size=15)\n",
    "ax.set_xlabel('Number of Interventions')\n",
    "    #ax.set_title('Change in False Positive Rate across different Interventions for COMPAS dataset', size=16)\n",
    "    #ax.set_title(title, size=16)\n",
    "ax.set_xticks(x, xlabels, rotation = 0, size=15)\n",
    "ax.set_ylim([0, 92])\n",
    "ax.legend(bbox_to_anchor=(0.5, 1.13), loc='upper center', ncol=2)\n",
    "\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "ax.spines['left'].set_visible(False)\n",
    "ax.spines['bottom'].set_color('#DDDDDD')\n",
    "#ax.tick_params(bottom=False, left=False)\n",
    "ax.set_axisbelow(True)\n",
    "ax.yaxis.grid(True, color='#EEEEEE')\n",
    "ax.xaxis.grid(False)\n",
    "\n",
    "ax.bar_label(rects1, padding=3)\n",
    "ax.bar_label(rects2, padding=3)\n",
    "\n",
    "fig.tight_layout()\n",
    "#plt.show()\n",
    "fig.savefig('./figures/accuracy_score.png', dpi=600)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}