cleanml-analysis-per-model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "73dff9bf",
   "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>model</th>\n",
       "      <th>error</th>\n",
       "      <th>fairness_impact</th>\n",
       "      <th>accuracy_impact</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>positive</td>\n",
       "      <td>positive</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>negative</td>\n",
       "      <td>positive</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>positive</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>positive</td>\n",
       "      <td>positive</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>positive</td>\n",
       "      <td>negative</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>negative</td>\n",
       "      <td>positive</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>negative</td>\n",
       "      <td>negative</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>positive</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>negative</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>positive</td>\n",
       "      <td>positive</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>positive</td>\n",
       "      <td>negative</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>positive</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>negative</td>\n",
       "      <td>positive</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>negative</td>\n",
       "      <td>negative</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>negative</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>positive</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>negative</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>51</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>positive</td>\n",
       "      <td>positive</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>negative</td>\n",
       "      <td>positive</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>negative</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>positive</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>positive</td>\n",
       "      <td>positive</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>positive</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>negative</td>\n",
       "      <td>positive</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>negative</td>\n",
       "      <td>negative</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>negative</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>positive</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>negative</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>outliers</td>\n",
       "      <td>positive</td>\n",
       "      <td>positive</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>outliers</td>\n",
       "      <td>positive</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>outliers</td>\n",
       "      <td>negative</td>\n",
       "      <td>positive</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>outliers</td>\n",
       "      <td>negative</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>outliers</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>positive</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>outliers</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>negative</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>outliers</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>positive</td>\n",
       "      <td>positive</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>negative</td>\n",
       "      <td>positive</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>mislabel</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>positive</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>positive</td>\n",
       "      <td>positive</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>positive</td>\n",
       "      <td>negative</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>positive</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>negative</td>\n",
       "      <td>positive</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>negative</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>positive</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>negative</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>positive</td>\n",
       "      <td>positive</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>positive</td>\n",
       "      <td>negative</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>positive</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>negative</td>\n",
       "      <td>positive</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>negative</td>\n",
       "      <td>negative</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>negative</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>positive</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>negative</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>insignificant</td>\n",
       "      <td>55</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  model           error fairness_impact accuracy_impact  count\n",
       "0               XGBoost        mislabel        positive        positive      4\n",
       "1               XGBoost        mislabel        negative        positive      4\n",
       "2               XGBoost        mislabel   insignificant        positive      4\n",
       "3               XGBoost        mislabel   insignificant   insignificant      2\n",
       "4               XGBoost  missing_values        positive        positive      8\n",
       "5               XGBoost  missing_values        positive        negative      2\n",
       "6               XGBoost  missing_values        negative        positive     10\n",
       "7               XGBoost  missing_values        negative        negative      3\n",
       "8               XGBoost  missing_values   insignificant        positive     40\n",
       "9               XGBoost  missing_values   insignificant        negative      3\n",
       "10              XGBoost  missing_values   insignificant   insignificant      6\n",
       "11              XGBoost        outliers        positive        positive      2\n",
       "12              XGBoost        outliers        positive        negative      5\n",
       "13              XGBoost        outliers        positive   insignificant      2\n",
       "14              XGBoost        outliers        negative        positive      4\n",
       "15              XGBoost        outliers        negative        negative     19\n",
       "16              XGBoost        outliers        negative   insignificant      5\n",
       "17              XGBoost        outliers   insignificant        positive     14\n",
       "18              XGBoost        outliers   insignificant        negative     24\n",
       "19              XGBoost        outliers   insignificant   insignificant     51\n",
       "20   knn_classification        mislabel        positive        positive      2\n",
       "21   knn_classification        mislabel        negative        positive      4\n",
       "22   knn_classification        mislabel        negative   insignificant      1\n",
       "23   knn_classification        mislabel   insignificant        positive      6\n",
       "24   knn_classification        mislabel   insignificant   insignificant      1\n",
       "25   knn_classification  missing_values        positive        positive     12\n",
       "26   knn_classification  missing_values        positive   insignificant      2\n",
       "27   knn_classification  missing_values        negative        positive      8\n",
       "28   knn_classification  missing_values        negative        negative      1\n",
       "29   knn_classification  missing_values        negative   insignificant     14\n",
       "30   knn_classification  missing_values   insignificant        positive      4\n",
       "31   knn_classification  missing_values   insignificant        negative     15\n",
       "32   knn_classification  missing_values   insignificant   insignificant     16\n",
       "33   knn_classification        outliers        positive        positive     11\n",
       "34   knn_classification        outliers        positive   insignificant      2\n",
       "35   knn_classification        outliers        negative        positive     18\n",
       "36   knn_classification        outliers        negative   insignificant      6\n",
       "37   knn_classification        outliers   insignificant        positive     31\n",
       "38   knn_classification        outliers   insignificant        negative      2\n",
       "39   knn_classification        outliers   insignificant   insignificant     56\n",
       "40  logistic_regression        mislabel        positive        positive      4\n",
       "41  logistic_regression        mislabel        negative        positive      5\n",
       "42  logistic_regression        mislabel   insignificant        positive      5\n",
       "43  logistic_regression  missing_values        positive        positive     10\n",
       "44  logistic_regression  missing_values        positive        negative      1\n",
       "45  logistic_regression  missing_values        positive   insignificant      2\n",
       "46  logistic_regression  missing_values        negative        positive      6\n",
       "47  logistic_regression  missing_values        negative   insignificant      9\n",
       "48  logistic_regression  missing_values   insignificant        positive      8\n",
       "49  logistic_regression  missing_values   insignificant        negative      3\n",
       "50  logistic_regression  missing_values   insignificant   insignificant     33\n",
       "51  logistic_regression        outliers        positive        positive      2\n",
       "52  logistic_regression        outliers        positive        negative      2\n",
       "53  logistic_regression        outliers        positive   insignificant      5\n",
       "54  logistic_regression        outliers        negative        positive      1\n",
       "55  logistic_regression        outliers        negative        negative     15\n",
       "56  logistic_regression        outliers        negative   insignificant      6\n",
       "57  logistic_regression        outliers   insignificant        positive     23\n",
       "58  logistic_regression        outliers   insignificant        negative     17\n",
       "59  logistic_regression        outliers   insignificant   insignificant     55"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import duckdb\n",
    "\n",
    "con = duckdb.connect(database=':memory:')\n",
    "con.execute(\"\"\"\n",
    "    SELECT model, error, fairness_impact, accuracy_impact, COUNT(*) as count\n",
    "    FROM 'cleanml.csv' \n",
    "    GROUP BY model, error, fairness_impact, accuracy_impact\n",
    "    ORDER BY model, error, fairness_impact DESC, accuracy_impact DESC\n",
    "\"\"\").df()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dc1fb2dc",
   "metadata": {},
   "source": [
    "## How often did data cleaning not negatively impacting fairness?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7d76efbd",
   "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>model</th>\n",
       "      <th>error</th>\n",
       "      <th>non_negative_impact</th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>True</td>\n",
       "      <td>134</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>False</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>True</td>\n",
       "      <td>147</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>outliers</td>\n",
       "      <td>False</td>\n",
       "      <td>69</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>True</td>\n",
       "      <td>66</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>False</td>\n",
       "      <td>78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>outliers</td>\n",
       "      <td>True</td>\n",
       "      <td>187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>outliers</td>\n",
       "      <td>False</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>True</td>\n",
       "      <td>86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>False</td>\n",
       "      <td>58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>True</td>\n",
       "      <td>137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>outliers</td>\n",
       "      <td>False</td>\n",
       "      <td>79</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  model           error  non_negative_impact  count\n",
       "0               XGBoost  missing_values                 True    134\n",
       "1               XGBoost  missing_values                False     10\n",
       "2               XGBoost        outliers                 True    147\n",
       "3               XGBoost        outliers                False     69\n",
       "4    knn_classification  missing_values                 True     66\n",
       "5    knn_classification  missing_values                False     78\n",
       "6    knn_classification        outliers                 True    187\n",
       "7    knn_classification        outliers                False     29\n",
       "8   logistic_regression  missing_values                 True     86\n",
       "9   logistic_regression  missing_values                False     58\n",
       "10  logistic_regression        outliers                 True    137\n",
       "11  logistic_regression        outliers                False     79"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "con.execute(\"\"\"\n",
    "    SELECT \n",
    "        model, error,    \n",
    "        fairness_impact!='negative' AND accuracy_impact!='negative' as non_negative_impact, \n",
    "        COUNT(*) as count    \n",
    "    FROM 'cleanml.csv' \n",
    "    GROUP BY model, error, non_negative_impact\n",
    "    ORDER BY model, error, non_negative_impact DESC\n",
    "\"\"\").df()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cba45027",
   "metadata": {},
   "source": [
    "## Is there a cleaning approach (per dataset, error and metric) to improve fairness?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "06953c9d",
   "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>model</th>\n",
       "      <th>dataset</th>\n",
       "      <th>error</th>\n",
       "      <th>metric</th>\n",
       "      <th>cleaning_found</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>adult</td>\n",
       "      <td>outliers</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>adult</td>\n",
       "      <td>outliers</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>folktables</td>\n",
       "      <td>outliers</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>german</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>folktables</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>credit</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>adult</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>credit</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>german</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>adult</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>credit</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>credit</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>adult</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  model     dataset           error             metric  \\\n",
       "0   logistic_regression       adult        outliers  equal_opportunity   \n",
       "1               XGBoost       adult        outliers  equal_opportunity   \n",
       "2               XGBoost  folktables        outliers  equal_opportunity   \n",
       "3   logistic_regression  folktables  missing_values  equal_opportunity   \n",
       "4    knn_classification  folktables  missing_values  equal_opportunity   \n",
       "5    knn_classification      german  missing_values  equal_opportunity   \n",
       "6               XGBoost  folktables  missing_values  equal_opportunity   \n",
       "7   logistic_regression  folktables        outliers  predictive_parity   \n",
       "8   logistic_regression      credit        outliers  predictive_parity   \n",
       "9    knn_classification       adult        outliers  predictive_parity   \n",
       "10   knn_classification      credit        outliers  predictive_parity   \n",
       "11   knn_classification      german        outliers  predictive_parity   \n",
       "12  logistic_regression       adult  missing_values  predictive_parity   \n",
       "13  logistic_regression  folktables  missing_values  predictive_parity   \n",
       "14  logistic_regression      credit  missing_values  predictive_parity   \n",
       "15   knn_classification  folktables  missing_values  predictive_parity   \n",
       "16   knn_classification      credit  missing_values  predictive_parity   \n",
       "17              XGBoost       adult  missing_values  predictive_parity   \n",
       "\n",
       "    cleaning_found  \n",
       "0             True  \n",
       "1             True  \n",
       "2             True  \n",
       "3             True  \n",
       "4             True  \n",
       "5             True  \n",
       "6             True  \n",
       "7             True  \n",
       "8             True  \n",
       "9             True  \n",
       "10            True  \n",
       "11            True  \n",
       "12            True  \n",
       "13            True  \n",
       "14            True  \n",
       "15            True  \n",
       "16            True  \n",
       "17            True  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "con.execute(\"\"\"\n",
    "    SELECT DISTINCT model, dataset, error, metric, TRUE as cleaning_found\n",
    "    FROM 'cleanml.csv' as csv\n",
    "    WHERE EXISTS(\n",
    "        SELECT model, dataset, error, metric FROM 'cleanml.csv' AS csv2\n",
    "        WHERE \n",
    "            csv2.model = csv.model AND\n",
    "            csv2.dataset = csv.dataset AND\n",
    "            csv2.error = csv.error AND\n",
    "            csv2.metric = csv.metric AND\n",
    "            csv2.fairness_impact = 'positive'\n",
    "    )    \n",
    "\"\"\").df()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a820ac1f",
   "metadata": {},
   "source": [
    "## Is there a cleaning approach (per dataset, error and metric) to improve fairness without decreasing accuracy?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e0750990",
   "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>model</th>\n",
       "      <th>dataset</th>\n",
       "      <th>error</th>\n",
       "      <th>metric</th>\n",
       "      <th>cleaning_found</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>adult</td>\n",
       "      <td>outliers</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>german</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>equal_opportunity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>folktables</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>credit</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>adult</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>credit</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>german</td>\n",
       "      <td>outliers</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>adult</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>logistic_regression</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>knn_classification</td>\n",
       "      <td>folktables</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>XGBoost</td>\n",
       "      <td>adult</td>\n",
       "      <td>missing_values</td>\n",
       "      <td>predictive_parity</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  model     dataset           error             metric  \\\n",
       "0               XGBoost       adult        outliers  equal_opportunity   \n",
       "1   logistic_regression  folktables  missing_values  equal_opportunity   \n",
       "2    knn_classification  folktables  missing_values  equal_opportunity   \n",
       "3    knn_classification      german  missing_values  equal_opportunity   \n",
       "4               XGBoost  folktables  missing_values  equal_opportunity   \n",
       "5   logistic_regression  folktables        outliers  predictive_parity   \n",
       "6   logistic_regression      credit        outliers  predictive_parity   \n",
       "7    knn_classification       adult        outliers  predictive_parity   \n",
       "8    knn_classification      credit        outliers  predictive_parity   \n",
       "9    knn_classification      german        outliers  predictive_parity   \n",
       "10  logistic_regression       adult  missing_values  predictive_parity   \n",
       "11  logistic_regression  folktables  missing_values  predictive_parity   \n",
       "12   knn_classification  folktables  missing_values  predictive_parity   \n",
       "13              XGBoost       adult  missing_values  predictive_parity   \n",
       "\n",
       "    cleaning_found  \n",
       "0             True  \n",
       "1             True  \n",
       "2             True  \n",
       "3             True  \n",
       "4             True  \n",
       "5             True  \n",
       "6             True  \n",
       "7             True  \n",
       "8             True  \n",
       "9             True  \n",
       "10            True  \n",
       "11            True  \n",
       "12            True  \n",
       "13            True  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "con.execute(\"\"\"\n",
    "    SELECT DISTINCT model, dataset, error, metric, TRUE as cleaning_found\n",
    "    FROM 'cleanml.csv' as csv\n",
    "    WHERE EXISTS(\n",
    "        SELECT model, dataset, error, metric FROM 'cleanml.csv' AS csv2\n",
    "        WHERE \n",
    "            csv2.model = csv.model AND\n",
    "            csv2.dataset = csv.dataset AND\n",
    "            csv2.error = csv.error AND\n",
    "            csv2.metric = csv.metric AND\n",
    "            csv2.fairness_impact = 'positive' AND csv2.accuracy_impact!='negative'\n",
    "    )   \n",
    "\"\"\").df()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "58469a54",
   "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}