04-Anomaly_Detection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Anomaly Detection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from src.setup import *\n",
    "%load_ext rpy2.ipython"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture --no-stdout err\n",
    "%%R\n",
    "library(tidyverse)\n",
    "require(\"RPostgreSQL\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Missing Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n=7:\n",
      "  method1: 1.109\n",
      "  method2: 1.109\n",
      "n=8:\n",
      "  method1: 1.229\n",
      "  method2: 1.229\n",
      "n=9:\n",
      "  method1: nan\n",
      "  method2: 1.510\n"
     ]
    }
   ],
   "source": [
    "for n in range(7, 10):\n",
    "    exp1 = 2**n\n",
    "    a = (22/7) ** exp1 \n",
    "    b = π ** exp1\n",
    "    # Compute answer in two \"equivalent\" ways\n",
    "    res1 = (a * a) / (b * b)\n",
    "    res2 = (a / b) * (a / b)\n",
    "    print(f\"n={n}:\\n  \"\n",
    "          f\"method1: {res1:.3f}\\n  \"\n",
    "          f\"method2: {res2:.3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### SQL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# PostgreSQL configuration\n",
    "def connect_local():\n",
    "    user = 'cleaning'\n",
    "    pwd = 'data'\n",
    "    host = 'localhost'\n",
    "    port = '5432'  \n",
    "    db = 'dirty'\n",
    "    con = psycopg2.connect(database=db, host=host, user=user, password=pwd)\n",
    "    engine = create_engine(f'postgresql://{user}:{pwd}@{host}:{port}/{db}')\n",
    "    return con, engine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 |  nan | Not number\n",
      "2 | 1.23 | A number  \n",
      "3 | None | A null    \n",
      "4 | 3.45 | Santiago  \n",
      "5 | 6.78 |           \n",
      "6 | 9.01 | None\n"
     ]
    }
   ],
   "source": [
    "con, engine = connect_local()\n",
    "cur = con.cursor()\n",
    "# Look at table named \"missing\"\n",
    "cur.execute(\"SELECT * FROM missing\")\n",
    "for n, (a, b) in enumerate(cur):\n",
    "    print(f\"{n+1} | {str(a):>4s} | {b}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[90m# A tibble: 3 x 2\u001b[39m\n",
      "    val str  \n",
      "  \u001b[3m\u001b[90m<dbl>\u001b[39m\u001b[23m \u001b[3m\u001b[90m<chr>\u001b[39m\u001b[23m\n",
      "\u001b[90m1\u001b[39m    \u001b[31mNA\u001b[39m this \n",
      "\u001b[90m2\u001b[39m   \u001b[31mNaN\u001b[39m that \n",
      "\u001b[90m3\u001b[39m     0 \u001b[31mNA\u001b[39m   \n"
     ]
    }
   ],
   "source": [
    "%%R\n",
    "# Notice NULL is simply ignored in the construction\n",
    "tibble(val = c(NULL, NA, NaN, 0), \n",
    "       str = c(\"this\", \"that\", NA))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[90m# A tibble: 6 x 2\u001b[39m\n",
      "       a b           \n",
      "   \u001b[3m\u001b[90m<dbl>\u001b[39m\u001b[23m \u001b[3m\u001b[90m<chr>\u001b[39m\u001b[23m       \n",
      "\u001b[90m1\u001b[39m \u001b[31mNaN\u001b[39m    \u001b[90m\"\u001b[39mNot number\u001b[90m\"\u001b[39m\n",
      "\u001b[90m2\u001b[39m   1.23 \u001b[90m\"\u001b[39mA number  \u001b[90m\"\u001b[39m\n",
      "\u001b[90m3\u001b[39m  \u001b[31mNA\u001b[39m    \u001b[90m\"\u001b[39mA null    \u001b[90m\"\u001b[39m\n",
      "\u001b[90m4\u001b[39m   3.45 \u001b[90m\"\u001b[39mSantiago  \u001b[90m\"\u001b[39m\n",
      "\u001b[90m5\u001b[39m   6.78 \u001b[90m\"\u001b[39m          \u001b[90m\"\u001b[39m\n",
      "\u001b[90m6\u001b[39m   9.01  \u001b[31mNA\u001b[39m         \n"
     ]
    }
   ],
   "source": [
    "%%R\n",
    "drv <- dbDriver(\"PostgreSQL\")\n",
    "con <- dbConnect(drv, dbname = \"dirty\",\n",
    "                 host = \"localhost\", port = 5432,\n",
    "                 user = \"cleaning\", password = \"data\")\n",
    "sql <- \"SELECT * FROM missing\"\n",
    "data <- tibble(dbGetQuery(con, sql))\n",
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NaN</td>\n",
       "      <td>Not number</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.23</td>\n",
       "      <td>A number</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>NaN</td>\n",
       "      <td>A null</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3.45</td>\n",
       "      <td>Santiago</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6.78</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>9.01</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      a           b\n",
       "0   NaN  Not number\n",
       "1  1.23  A number  \n",
       "2   NaN  A null    \n",
       "3  3.45  Santiago  \n",
       "4  6.78            \n",
       "5  9.01        None"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.read_sql(\"SELECT * FROM missing\", engine)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Hierarchical Formats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[nan, None, inf]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "json.loads('[NaN, null, Infinity]')  # null becomes Python None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "json_data = '''\n",
    "{\"a\": {\"1\": NaN, \"2\": 1.23, \"4\": 3.45, \"5\": 6.78, \"6\": 9.01},\n",
    " \"b\": {\"1\": \"Not number\", \"2\": \"A number\", \"3\": \"A null\",\n",
    "       \"4\": \"Santiago\", \"5\": \"\"}\n",
    "}'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>a</th>\n",
       "      <th>b</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NaN</td>\n",
       "      <td>Not number</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.23</td>\n",
       "      <td>A number</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>NaN</td>\n",
       "      <td>A null</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.45</td>\n",
       "      <td>Santiago</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6.78</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>9.01</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      a           b\n",
       "1   NaN  Not number\n",
       "2  1.23    A number\n",
       "3   NaN      A null\n",
       "4  3.45    Santiago\n",
       "5  6.78            \n",
       "6  9.01         NaN"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.read_json(json_data).sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Row 1, Col a: Not a Number\n",
      "Row 3, Col a: Missing\n",
      "Row 5, Col b: Empty value ''\n",
      "Row 6, Col b: Missing\n"
     ]
    }
   ],
   "source": [
    "data = json.loads(json_data)\n",
    "rows = {row for dct in data.values() \n",
    "            for row in dct.keys()}\n",
    "\n",
    "for row in sorted(rows):\n",
    "    for col in data.keys():\n",
    "        val = data[col].get(row)\n",
    "        if val is None:\n",
    "            print(f\"Row {row}, Col {col}: Missing\")\n",
    "        elif isinstance(val, float) and math.isnan(val):\n",
    "            print(f\"Row {row}, Col {col}: Not a Number\")\n",
    "        elif not val:\n",
    "            print(f\"Row {row}, Col {col}: Empty value {repr(val)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sentinels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>STATION</th>\n",
       "      <th>DATE</th>\n",
       "      <th>TEMP</th>\n",
       "      <th>VISIB</th>\n",
       "      <th>GUST</th>\n",
       "      <th>DEWP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-01-01</td>\n",
       "      <td>39.7</td>\n",
       "      <td>6.2</td>\n",
       "      <td>52.1</td>\n",
       "      <td>30.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-01-02</td>\n",
       "      <td>36.4</td>\n",
       "      <td>6.2</td>\n",
       "      <td>999.9</td>\n",
       "      <td>29.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-01-03</td>\n",
       "      <td>36.5</td>\n",
       "      <td>3.3</td>\n",
       "      <td>999.9</td>\n",
       "      <td>35.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>UNKNOWN</td>\n",
       "      <td>45.6</td>\n",
       "      <td>2.2</td>\n",
       "      <td>22.0</td>\n",
       "      <td>44.8</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>295</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-12-17</td>\n",
       "      <td>40.5</td>\n",
       "      <td>6.2</td>\n",
       "      <td>999.9</td>\n",
       "      <td>39.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-12-18</td>\n",
       "      <td>38.8</td>\n",
       "      <td>6.2</td>\n",
       "      <td>999.9</td>\n",
       "      <td>38.2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>297</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-12-19</td>\n",
       "      <td>45.5</td>\n",
       "      <td>6.1</td>\n",
       "      <td>999.9</td>\n",
       "      <td>42.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>298</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-12-20</td>\n",
       "      <td>51.8</td>\n",
       "      <td>6.2</td>\n",
       "      <td>35.0</td>\n",
       "      <td>41.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>299 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        STATION        DATE  TEMP  VISIB   GUST  DEWP\n",
       "0    1001499999  2019-01-01  39.7    6.2   52.1  30.4\n",
       "1    1001499999  2019-01-02  36.4    6.2  999.9  29.8\n",
       "2    1001499999  2019-01-03  36.5    3.3  999.9  35.6\n",
       "3    1001499999     UNKNOWN  45.6    2.2   22.0  44.8\n",
       "..          ...         ...   ...    ...    ...   ...\n",
       "295  1001499999  2019-12-17  40.5    6.2  999.9  39.2\n",
       "296  1001499999  2019-12-18  38.8    6.2  999.9  38.2\n",
       "297  1001499999  2019-12-19  45.5    6.1  999.9  42.7\n",
       "298  1001499999  2019-12-20  51.8    6.2   35.0  41.2\n",
       "\n",
       "[299 rows x 6 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sorstokken = pd.read_csv('data/sorstokken-no.csv.gz')\n",
    "sorstokken"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>STATION</th>\n",
       "      <th>DATE</th>\n",
       "      <th>TEMP</th>\n",
       "      <th>VISIB</th>\n",
       "      <th>GUST</th>\n",
       "      <th>DEWP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-01-01</td>\n",
       "      <td>27.2</td>\n",
       "      <td>1.2</td>\n",
       "      <td>17.1</td>\n",
       "      <td>16.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>UNKNOWN</td>\n",
       "      <td>88.1</td>\n",
       "      <td>999.9</td>\n",
       "      <td>999.9</td>\n",
       "      <td>63.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      STATION        DATE  TEMP  VISIB   GUST  DEWP\n",
       "0  1001499999  2019-01-01  27.2    1.2   17.1  16.5\n",
       "1  1001499999     UNKNOWN  88.1  999.9  999.9  63.5"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame([sorstokken.min(), sorstokken.max()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Normal max:\n",
      "VISIB 6.8 ...standard deviation w/ & w/o sentinel: 254.4 / 0.7\n",
      "GUST 62.2 ...standard deviation w/ & w/o sentinel: 452.4 / 8.1\n"
     ]
    }
   ],
   "source": [
    "print(\"Normal max:\")\n",
    "for col in ['VISIB', 'GUST']:\n",
    "        s = sorstokken[col]\n",
    "        print(col, s[s < 999.9].max(), \n",
    "              \"...standard deviation w/ & w/o sentinel:\",\n",
    "              f\"{s.std():.1f} / {s[s < 999.9].std():.1f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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>STATION</th>\n",
       "      <th>DATE</th>\n",
       "      <th>TEMP</th>\n",
       "      <th>VISIB</th>\n",
       "      <th>GUST</th>\n",
       "      <th>DEWP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-01-01</td>\n",
       "      <td>39.7</td>\n",
       "      <td>6.2</td>\n",
       "      <td>52.1</td>\n",
       "      <td>30.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-01-02</td>\n",
       "      <td>36.4</td>\n",
       "      <td>6.2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>29.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-01-03</td>\n",
       "      <td>36.5</td>\n",
       "      <td>3.3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>35.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>NaT</td>\n",
       "      <td>45.6</td>\n",
       "      <td>2.2</td>\n",
       "      <td>22.0</td>\n",
       "      <td>44.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1001499999</td>\n",
       "      <td>2019-01-06</td>\n",
       "      <td>42.5</td>\n",
       "      <td>1.9</td>\n",
       "      <td>NaN</td>\n",
       "      <td>42.5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      STATION       DATE  TEMP  VISIB  GUST  DEWP\n",
       "0  1001499999 2019-01-01  39.7    6.2  52.1  30.4\n",
       "1  1001499999 2019-01-02  36.4    6.2   NaN  29.8\n",
       "2  1001499999 2019-01-03  36.5    3.3   NaN  35.6\n",
       "3  1001499999        NaT  45.6    2.2  22.0  44.8\n",
       "4  1001499999 2019-01-06  42.5    1.9   NaN  42.5"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sorstokken = pd.read_csv('data/sorstokken-no.csv.gz', \n",
    "                         na_values={'DATE': 'UNKNOWN', \n",
    "                                    'VISIB': '999.9',\n",
    "                                    'GUST': '999.9'},\n",
    "                         parse_dates=['DATE'])\n",
    "sorstokken.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Miscoded Data"
   ]
  },
  {
   "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>min</th>\n",
       "      <th>max</th>\n",
       "      <th>dtype</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>erythema</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>int64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scaling</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>int64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>definite borders</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>int64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>itching</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>int64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>inflammatory monoluclear inflitrate</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>int64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>band-like infiltrate</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>int64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Age</th>\n",
       "      <td>0</td>\n",
       "      <td>75</td>\n",
       "      <td>float64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TARGET</th>\n",
       "      <td>cronic dermatitis</td>\n",
       "      <td>seboreic dermatitis</td>\n",
       "      <td>object</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>35 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   min                  max  \\\n",
       "erythema                                             0                    3   \n",
       "scaling                                              0                    3   \n",
       "definite borders                                     0                    3   \n",
       "itching                                              0                    3   \n",
       "...                                                ...                  ...   \n",
       "inflammatory monoluclear inflitrate                  0                    3   \n",
       "band-like infiltrate                                 0                    3   \n",
       "Age                                                  0                   75   \n",
       "TARGET                               cronic dermatitis  seboreic dermatitis   \n",
       "\n",
       "                                       dtype  \n",
       "erythema                               int64  \n",
       "scaling                                int64  \n",
       "definite borders                       int64  \n",
       "itching                                int64  \n",
       "...                                      ...  \n",
       "inflammatory monoluclear inflitrate    int64  \n",
       "band-like infiltrate                   int64  \n",
       "Age                                  float64  \n",
       "TARGET                                object  \n",
       "\n",
       "[35 rows x 3 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from src.dermatology import *\n",
    "(pd.DataFrame(\n",
    "    [derm.min(), derm.max(), derm.dtypes])\n",
    "     .T\n",
    "     .rename(columns={0:'min', 1:'max', 2:'dtype'})\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Hair_Length</th>\n",
       "      <th>Favorite</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21492</th>\n",
       "      <td>176.958650</td>\n",
       "      <td>72.604585</td>\n",
       "      <td>14.0</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9488</th>\n",
       "      <td>169.000221</td>\n",
       "      <td>79.559843</td>\n",
       "      <td>0.0</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16933</th>\n",
       "      <td>171.104306</td>\n",
       "      <td>71.125528</td>\n",
       "      <td>5.5</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12604</th>\n",
       "      <td>174.481084</td>\n",
       "      <td>79.496237</td>\n",
       "      <td>8.1</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8222</th>\n",
       "      <td>171.275578</td>\n",
       "      <td>77.094118</td>\n",
       "      <td>14.6</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight  Hair_Length Favorite\n",
       "21492  176.958650  72.604585         14.0      red\n",
       "9488   169.000221  79.559843          0.0     blue\n",
       "16933  171.104306  71.125528          5.5      red\n",
       "12604  174.481084  79.496237          8.1     blue\n",
       "8222   171.275578  77.094118         14.6    green"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans = pd.read_csv('data/humans-err.csv')\n",
    "# random_state for deterministic sample\n",
    "humans.sample(5, random_state=1)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['red', 'green', 'blue', 'Red', ' red', 'grееn', 'blüe',\n",
       "       'chartreuse'], dtype=object)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans.Favorite.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "red           9576\n",
       "blue          7961\n",
       "green         7458\n",
       "grееn            1\n",
       " red             1\n",
       "blüe             1\n",
       "chartreuse       1\n",
       "Red              1\n",
       "Name: Favorite, dtype: int64"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans.Favorite.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "       red [32, 114, 101, 100]\n",
      "       Red [82, 101, 100]\n",
      "      blue [98, 108, 117, 101]\n",
      "      blüe [98, 108, 252, 101]\n",
      "chartreuse [99, 104, 97, 114, 116, 114, 101, 117, 115, 101]\n",
      "     green [103, 114, 101, 101, 110]\n",
      "     grееn [103, 114, 1077, 1077, 110]\n",
      "       red [114, 101, 100]\n"
     ]
    }
   ],
   "source": [
    "for color in sorted(humans.Favorite.unique()):\n",
    "    print(f\"{color:>10s}\", [ord(c) for c in color])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "red      9578\n",
       "blue     7962\n",
       "green    7459\n",
       "Name: Favorite, dtype: int64"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans.loc[humans.Favorite.isin(['Red', ' red']), 'Favorite'] = 'red'\n",
    "humans.loc[humans.Favorite == 'chartreuse', 'Favorite'] = None\n",
    "humans.loc[humans.Favorite == 'blüe', 'Favorite'] = 'blue'\n",
    "humans.loc[humans.Favorite == 'grееn', 'Favorite'] = 'green'\n",
    "humans.Favorite.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fixed Bounds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "((humans.Height < 92) | (humans.Height > 213)).any()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Hair_Length</th>\n",
       "      <th>Favorite</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1984</th>\n",
       "      <td>165.634695</td>\n",
       "      <td>62.979993</td>\n",
       "      <td>127.0</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8929</th>\n",
       "      <td>175.186061</td>\n",
       "      <td>73.899992</td>\n",
       "      <td>120.6</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14673</th>\n",
       "      <td>174.948037</td>\n",
       "      <td>77.644434</td>\n",
       "      <td>130.1</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14735</th>\n",
       "      <td>176.385525</td>\n",
       "      <td>68.735397</td>\n",
       "      <td>121.7</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16672</th>\n",
       "      <td>173.172298</td>\n",
       "      <td>71.814699</td>\n",
       "      <td>121.4</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17093</th>\n",
       "      <td>169.771111</td>\n",
       "      <td>77.958278</td>\n",
       "      <td>133.2</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight  Hair_Length Favorite\n",
       "1984   165.634695  62.979993        127.0      red\n",
       "8929   175.186061  73.899992        120.6     blue\n",
       "14673  174.948037  77.644434        130.1     blue\n",
       "14735  176.385525  68.735397        121.7    green\n",
       "16672  173.172298  71.814699        121.4      red\n",
       "17093  169.771111  77.958278        133.2     blue"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans.query('Hair_Length > 120')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Hair_Length</th>\n",
       "      <th>Favorite</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1984</th>\n",
       "      <td>165.634695</td>\n",
       "      <td>62.979993</td>\n",
       "      <td>120.0</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4146</th>\n",
       "      <td>173.930107</td>\n",
       "      <td>72.701456</td>\n",
       "      <td>119.6</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8929</th>\n",
       "      <td>175.186061</td>\n",
       "      <td>73.899992</td>\n",
       "      <td>120.0</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9259</th>\n",
       "      <td>179.215974</td>\n",
       "      <td>82.538890</td>\n",
       "      <td>119.4</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14673</th>\n",
       "      <td>174.948037</td>\n",
       "      <td>77.644434</td>\n",
       "      <td>120.0</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14735</th>\n",
       "      <td>176.385525</td>\n",
       "      <td>68.735397</td>\n",
       "      <td>120.0</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16672</th>\n",
       "      <td>173.172298</td>\n",
       "      <td>71.814699</td>\n",
       "      <td>120.0</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17093</th>\n",
       "      <td>169.771111</td>\n",
       "      <td>77.958278</td>\n",
       "      <td>120.0</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight  Hair_Length Favorite\n",
       "1984   165.634695  62.979993        120.0      red\n",
       "4146   173.930107  72.701456        119.6      red\n",
       "8929   175.186061  73.899992        120.0     blue\n",
       "9259   179.215974  82.538890        119.4    green\n",
       "14673  174.948037  77.644434        120.0     blue\n",
       "14735  176.385525  68.735397        120.0    green\n",
       "16672  173.172298  71.814699        120.0      red\n",
       "17093  169.771111  77.958278        120.0     blue"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans2 = humans.copy()  # Retain prior versions of data set\n",
    "humans2['Hair_Length'] = humans2.Hair_Length.clip(upper=120)\n",
    "humans2[humans2.Hair_Length > 119]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Hair_Length</th>\n",
       "      <th>Favorite</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>177.297182</td>\n",
       "      <td>81.153493</td>\n",
       "      <td>0.0</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>217</th>\n",
       "      <td>171.893967</td>\n",
       "      <td>68.553526</td>\n",
       "      <td>0.0</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>240</th>\n",
       "      <td>161.862237</td>\n",
       "      <td>76.914599</td>\n",
       "      <td>0.0</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>354</th>\n",
       "      <td>172.972247</td>\n",
       "      <td>73.175032</td>\n",
       "      <td>0.0</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24834</th>\n",
       "      <td>170.991301</td>\n",
       "      <td>67.652660</td>\n",
       "      <td>0.0</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24892</th>\n",
       "      <td>177.002643</td>\n",
       "      <td>77.286141</td>\n",
       "      <td>0.0</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24919</th>\n",
       "      <td>169.012286</td>\n",
       "      <td>74.593809</td>\n",
       "      <td>0.0</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24967</th>\n",
       "      <td>169.061308</td>\n",
       "      <td>65.985481</td>\n",
       "      <td>0.0</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>517 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight  Hair_Length Favorite\n",
       "6      177.297182  81.153493          0.0     blue\n",
       "217    171.893967  68.553526          0.0     blue\n",
       "240    161.862237  76.914599          0.0     blue\n",
       "354    172.972247  73.175032          0.0      red\n",
       "...           ...        ...          ...      ...\n",
       "24834  170.991301  67.652660          0.0    green\n",
       "24892  177.002643  77.286141          0.0    green\n",
       "24919  169.012286  74.593809          0.0     blue\n",
       "24967  169.061308  65.985481          0.0    green\n",
       "\n",
       "[517 rows x 4 columns]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans2[humans2.Hair_Length == 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Hair_Length</th>\n",
       "      <th>Favorite</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>493</th>\n",
       "      <td>167.703398</td>\n",
       "      <td>72.567763</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>blue</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>528</th>\n",
       "      <td>167.355393</td>\n",
       "      <td>60.276190</td>\n",
       "      <td>-20.7</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>562</th>\n",
       "      <td>172.416114</td>\n",
       "      <td>60.867457</td>\n",
       "      <td>-68.1</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>569</th>\n",
       "      <td>177.644146</td>\n",
       "      <td>74.027147</td>\n",
       "      <td>-5.9</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24055</th>\n",
       "      <td>172.831608</td>\n",
       "      <td>74.096660</td>\n",
       "      <td>-13.3</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24063</th>\n",
       "      <td>172.687488</td>\n",
       "      <td>69.466838</td>\n",
       "      <td>-14.2</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24386</th>\n",
       "      <td>176.668430</td>\n",
       "      <td>62.984811</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>green</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24944</th>\n",
       "      <td>172.300925</td>\n",
       "      <td>72.067862</td>\n",
       "      <td>-24.4</td>\n",
       "      <td>red</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>118 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight  Hair_Length Favorite\n",
       "493    167.703398  72.567763         -1.0     blue\n",
       "528    167.355393  60.276190        -20.7    green\n",
       "562    172.416114  60.867457        -68.1    green\n",
       "569    177.644146  74.027147         -5.9    green\n",
       "...           ...        ...          ...      ...\n",
       "24055  172.831608  74.096660        -13.3      red\n",
       "24063  172.687488  69.466838        -14.2    green\n",
       "24386  176.668430  62.984811         -1.0    green\n",
       "24944  172.300925  72.067862        -24.4      red\n",
       "\n",
       "[118 rows x 4 columns]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "neg_hair = humans2[humans2.Hair_Length < 0]\n",
    "neg_hair"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. \n",
      " -95.70  -38.08  -20.65  -24.35   -5.60   -0.70 \n"
     ]
    }
   ],
   "source": [
    "%%R -i neg_hair\n",
    "summary(neg_hair$Hair_Length)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    118.000000\n",
       "mean     -24.348305\n",
       "std       22.484691\n",
       "min      -95.700000\n",
       "25%      -38.075000\n",
       "50%      -20.650000\n",
       "75%       -5.600000\n",
       "max       -0.700000\n",
       "Name: Hair_Length, dtype: float64"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "neg_hair.Hair_Length.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x400 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(neg_hair.Hair_Length, bins=30)\n",
    "plt.title(\"Distribution of invalid negative hair length\")\n",
    "plt.savefig(\"img/(Ch04)Distribution of invalid negative hair length.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-1.0     19\n",
       "-41.6     2\n",
       "-6.8      2\n",
       "-30.1     2\n",
       "         ..\n",
       "-3.3      1\n",
       "-51.4     1\n",
       "-25.1     1\n",
       "-4.8      1\n",
       "Name: Hair_Length, Length: 93, dtype: int64"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "neg_hair.Hair_Length.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x400 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "humans3 = humans2.copy()     # Versioned changes to data\n",
    "\n",
    "# The \"sentinel\" negative value means missing\n",
    "humans3.loc[humans3.Hair_Length == -1, 'Hair_Length'] = None\n",
    "\n",
    "# All other values simply become non-negative\n",
    "humans3['Hair_Length'] = humans3.Hair_Length.abs()\n",
    "\n",
    "plt.hist(humans3.Hair_Length, bins=30)\n",
    "plt.title(\"Distribution of corrected hair lengths\")\n",
    "plt.savefig(\"img/(Ch04)Distribution of corrected hair lengths.png\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Outliers "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Z-Score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 4))\n",
    "ax1.hist(humans3.Height, bins=50)\n",
    "ax2.hist(humans3.Weight, bins=50)\n",
    "ax1.set_title(\"Distribution of Height\")\n",
    "ax2.set_title(\"Distribution of Weight\")\n",
    "fig.savefig(\"img/(Ch04)Distribution of Height and Weight.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Anderson-Darling statistic for Height      :   0.24\n",
      "Anderson-Darling statistic for Weight      :   0.54\n",
      "Anderson-Darling statistic for Hair_Length : 578.19\n"
     ]
    }
   ],
   "source": [
    "from scipy.stats import anderson\n",
    "\n",
    "for var in ('Height', 'Weight', 'Hair_Length'):\n",
    "    data = humans3[var][humans3[var].notnull()]\n",
    "    stat = anderson(data, 'norm').statistic\n",
    "    print(f\"Anderson-Darling statistic for {var:<12s}: {stat:6.2f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Hair_Length</th>\n",
       "      <th>Favorite</th>\n",
       "      <th>zscore_Height</th>\n",
       "      <th>zscore_Weight</th>\n",
       "      <th>zscore_Hair_Length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21492</th>\n",
       "      <td>176.958650</td>\n",
       "      <td>72.604585</td>\n",
       "      <td>14.0</td>\n",
       "      <td>red</td>\n",
       "      <td>0.880831</td>\n",
       "      <td>-0.042032</td>\n",
       "      <td>-0.568786</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9488</th>\n",
       "      <td>169.000221</td>\n",
       "      <td>79.559843</td>\n",
       "      <td>0.0</td>\n",
       "      <td>blue</td>\n",
       "      <td>-0.766210</td>\n",
       "      <td>0.997585</td>\n",
       "      <td>-1.225152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16933</th>\n",
       "      <td>171.104306</td>\n",
       "      <td>71.125528</td>\n",
       "      <td>5.5</td>\n",
       "      <td>red</td>\n",
       "      <td>-0.330758</td>\n",
       "      <td>-0.263109</td>\n",
       "      <td>-0.967294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12604</th>\n",
       "      <td>174.481084</td>\n",
       "      <td>79.496237</td>\n",
       "      <td>8.1</td>\n",
       "      <td>blue</td>\n",
       "      <td>0.368085</td>\n",
       "      <td>0.988078</td>\n",
       "      <td>-0.845397</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8222</th>\n",
       "      <td>171.275578</td>\n",
       "      <td>77.094118</td>\n",
       "      <td>14.6</td>\n",
       "      <td>green</td>\n",
       "      <td>-0.295312</td>\n",
       "      <td>0.629028</td>\n",
       "      <td>-0.540656</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight  Hair_Length Favorite  zscore_Height  \\\n",
       "21492  176.958650  72.604585         14.0      red       0.880831   \n",
       "9488   169.000221  79.559843          0.0     blue      -0.766210   \n",
       "16933  171.104306  71.125528          5.5      red      -0.330758   \n",
       "12604  174.481084  79.496237          8.1     blue       0.368085   \n",
       "8222   171.275578  77.094118         14.6    green      -0.295312   \n",
       "\n",
       "       zscore_Weight  zscore_Hair_Length  \n",
       "21492      -0.042032           -0.568786  \n",
       "9488        0.997585           -1.225152  \n",
       "16933      -0.263109           -0.967294  \n",
       "12604       0.988078           -0.845397  \n",
       "8222        0.629028           -0.540656  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans4 = humans3.copy()\n",
    "\n",
    "for var in ('Height', 'Weight', 'Hair_Length'):\n",
    "    zscore = (humans4[var] - humans4[var].mean()) / humans4[var].std()\n",
    "    humans4[f\"zscore_{var}\"] = zscore\n",
    "    \n",
    "humans4.sample(5, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Hair_Length</th>\n",
       "      <th>Favorite</th>\n",
       "      <th>zscore_Height</th>\n",
       "      <th>zscore_Weight</th>\n",
       "      <th>zscore_Hair_Length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>138</th>\n",
       "      <td>187.708718</td>\n",
       "      <td>86.829633</td>\n",
       "      <td>19.3</td>\n",
       "      <td>green</td>\n",
       "      <td>3.105616</td>\n",
       "      <td>2.084216</td>\n",
       "      <td>-0.320304</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>174</th>\n",
       "      <td>187.537446</td>\n",
       "      <td>79.893761</td>\n",
       "      <td>37.5</td>\n",
       "      <td>blue</td>\n",
       "      <td>3.070170</td>\n",
       "      <td>1.047496</td>\n",
       "      <td>0.532971</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>412</th>\n",
       "      <td>157.522316</td>\n",
       "      <td>62.564977</td>\n",
       "      <td>6.8</td>\n",
       "      <td>blue</td>\n",
       "      <td>-3.141625</td>\n",
       "      <td>-1.542673</td>\n",
       "      <td>-0.906345</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1162</th>\n",
       "      <td>188.592435</td>\n",
       "      <td>86.155948</td>\n",
       "      <td>53.1</td>\n",
       "      <td>red</td>\n",
       "      <td>3.288506</td>\n",
       "      <td>1.983518</td>\n",
       "      <td>1.264351</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>22945</th>\n",
       "      <td>157.293031</td>\n",
       "      <td>44.744929</td>\n",
       "      <td>18.4</td>\n",
       "      <td>red</td>\n",
       "      <td>-3.189077</td>\n",
       "      <td>-4.206272</td>\n",
       "      <td>-0.362499</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23039</th>\n",
       "      <td>187.845548</td>\n",
       "      <td>88.554510</td>\n",
       "      <td>6.9</td>\n",
       "      <td>blue</td>\n",
       "      <td>3.133934</td>\n",
       "      <td>2.342037</td>\n",
       "      <td>-0.901657</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24244</th>\n",
       "      <td>158.153049</td>\n",
       "      <td>59.725932</td>\n",
       "      <td>13.8</td>\n",
       "      <td>green</td>\n",
       "      <td>-3.011091</td>\n",
       "      <td>-1.967031</td>\n",
       "      <td>-0.578162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24801</th>\n",
       "      <td>189.310696</td>\n",
       "      <td>85.406727</td>\n",
       "      <td>2.3</td>\n",
       "      <td>green</td>\n",
       "      <td>3.437154</td>\n",
       "      <td>1.871531</td>\n",
       "      <td>-1.117320</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>51 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight  Hair_Length Favorite  zscore_Height  \\\n",
       "138    187.708718  86.829633         19.3    green       3.105616   \n",
       "174    187.537446  79.893761         37.5     blue       3.070170   \n",
       "412    157.522316  62.564977          6.8     blue      -3.141625   \n",
       "1162   188.592435  86.155948         53.1      red       3.288506   \n",
       "...           ...        ...          ...      ...            ...   \n",
       "22945  157.293031  44.744929         18.4      red      -3.189077   \n",
       "23039  187.845548  88.554510          6.9     blue       3.133934   \n",
       "24244  158.153049  59.725932         13.8    green      -3.011091   \n",
       "24801  189.310696  85.406727          2.3    green       3.437154   \n",
       "\n",
       "       zscore_Weight  zscore_Hair_Length  \n",
       "138         2.084216           -0.320304  \n",
       "174         1.047496            0.532971  \n",
       "412        -1.542673           -0.906345  \n",
       "1162        1.983518            1.264351  \n",
       "...              ...                 ...  \n",
       "22945      -4.206272           -0.362499  \n",
       "23039       2.342037           -0.901657  \n",
       "24244      -1.967031           -0.578162  \n",
       "24801       1.871531           -1.117320  \n",
       "\n",
       "[51 rows x 7 columns]"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans4[humans4.zscore_Height.abs() > 3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Outlier weight: 67\n",
      "Outlier hair length: 285\n"
     ]
    }
   ],
   "source": [
    "print(\"Outlier weight:\", (humans4.zscore_Weight.abs() > 3).sum())\n",
    "print(\"Outlier hair length:\", (humans4.zscore_Hair_Length.abs() > 3).sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Hair_Length</th>\n",
       "      <th>Favorite</th>\n",
       "      <th>zscore_Height</th>\n",
       "      <th>zscore_Weight</th>\n",
       "      <th>zscore_Hair_Length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13971</th>\n",
       "      <td>153.107034</td>\n",
       "      <td>63.155154</td>\n",
       "      <td>4.4</td>\n",
       "      <td>green</td>\n",
       "      <td>-4.055392</td>\n",
       "      <td>-1.454458</td>\n",
       "      <td>-1.018865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14106</th>\n",
       "      <td>157.244415</td>\n",
       "      <td>45.062151</td>\n",
       "      <td>70.7</td>\n",
       "      <td>red</td>\n",
       "      <td>-3.199138</td>\n",
       "      <td>-4.158856</td>\n",
       "      <td>2.089496</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22945</th>\n",
       "      <td>157.293031</td>\n",
       "      <td>44.744929</td>\n",
       "      <td>18.4</td>\n",
       "      <td>red</td>\n",
       "      <td>-3.189077</td>\n",
       "      <td>-4.206272</td>\n",
       "      <td>-0.362499</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight  Hair_Length Favorite  zscore_Height  \\\n",
       "13971  153.107034  63.155154          4.4    green      -4.055392   \n",
       "14106  157.244415  45.062151         70.7      red      -3.199138   \n",
       "22945  157.293031  44.744929         18.4      red      -3.189077   \n",
       "\n",
       "       zscore_Weight  zscore_Hair_Length  \n",
       "13971      -1.454458           -1.018865  \n",
       "14106      -4.158856            2.089496  \n",
       "22945      -4.206272           -0.362499  "
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cond = (\n",
    "    (humans4.zscore_Height.abs() > 4) |\n",
    "    (humans4.zscore_Weight.abs() > 4) |\n",
    "    (humans4.zscore_Hair_Length.abs() > 4.5))\n",
    "humans4[cond] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Interquartile Range"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function defined but not run in this cell\n",
    "def show_boxplots(df, cols, whis=1.5):\n",
    "    # Create as many horizontal plots as we have columns\n",
    "    fig, axes = plt.subplots(len(cols), 1, figsize=(10, 2*len(cols)))\n",
    "    # For each one, plot the non-null data inside it\n",
    "    for n, col in enumerate(cols):\n",
    "        data = df[col][df[col].notnull()]\n",
    "        axes[n].set_title(f'{col} Distribution')\n",
    "        # Extend whiskers to specified IQR multiplier\n",
    "        axes[n].boxplot(data, whis=whis, vert=False, sym='x')\n",
    "        axes[n].set_yticks([])\n",
    "    # Fix spacing of subplots at the end\n",
    "    fig.tight_layout()\n",
    "    plt.savefig(f\"img/boxplot-{'_'.join(cols)}.png\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x600 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_boxplots(humans4, [\"Height\", \"Weight\", \"Hair_Length\"], 2.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0.25</th>\n",
       "      <td>169.428884</td>\n",
       "      <td>68.428823</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.50</th>\n",
       "      <td>172.709078</td>\n",
       "      <td>72.930616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0.75</th>\n",
       "      <td>175.953541</td>\n",
       "      <td>77.367039</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1.00</th>\n",
       "      <td>190.888112</td>\n",
       "      <td>98.032504</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Height     Weight\n",
       "0.25  169.428884  68.428823\n",
       "0.50  172.709078  72.930616\n",
       "0.75  175.953541  77.367039\n",
       "1.00  190.888112  98.032504"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "quartiles = (\n",
    "    humans4[['Height', 'Weight']]\n",
    "    .quantile(q=[0.25, 0.50, 0.75, 1.0]))\n",
    "quartiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Height    6.524657\n",
       "Weight    8.938216\n",
       "dtype: float64"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "IQR = quartiles.loc[0.75] - quartiles.loc[0.25]\n",
    "IQR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inliers for Height: [153.117, 192.265]\n",
      "Inliers for Weight: [46.083, 99.713]\n"
     ]
    }
   ],
   "source": [
    "for col, length in IQR.iteritems():\n",
    "    high = quartiles.loc[0.75, col] + 2.5*IQR[col]\n",
    "    low = quartiles.loc[0.25, col] - 2.5*IQR[col]\n",
    "    print(f\"Inliers for {col}: [{low:.3f}, {high:.3f}]\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>Hair_Length</th>\n",
       "      <th>Favorite</th>\n",
       "      <th>zscore_Height</th>\n",
       "      <th>zscore_Weight</th>\n",
       "      <th>zscore_Hair_Length</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13971</th>\n",
       "      <td>153.107034</td>\n",
       "      <td>63.155154</td>\n",
       "      <td>4.4</td>\n",
       "      <td>green</td>\n",
       "      <td>-4.055392</td>\n",
       "      <td>-1.454458</td>\n",
       "      <td>-1.018865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14106</th>\n",
       "      <td>157.244415</td>\n",
       "      <td>45.062151</td>\n",
       "      <td>70.7</td>\n",
       "      <td>red</td>\n",
       "      <td>-3.199138</td>\n",
       "      <td>-4.158856</td>\n",
       "      <td>2.089496</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22945</th>\n",
       "      <td>157.293031</td>\n",
       "      <td>44.744929</td>\n",
       "      <td>18.4</td>\n",
       "      <td>red</td>\n",
       "      <td>-3.189077</td>\n",
       "      <td>-4.206272</td>\n",
       "      <td>-0.362499</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight  Hair_Length Favorite  zscore_Height  \\\n",
       "13971  153.107034  63.155154          4.4    green      -4.055392   \n",
       "14106  157.244415  45.062151         70.7      red      -3.199138   \n",
       "22945  157.293031  44.744929         18.4      red      -3.189077   \n",
       "\n",
       "       zscore_Weight  zscore_Hair_Length  \n",
       "13971      -1.454458           -1.018865  \n",
       "14106      -4.158856            2.089496  \n",
       "22945      -4.206272           -0.362499  "
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cond = (\n",
    "    (humans4.Height > 192.265) |\n",
    "    (humans4.Height < 153.117) |\n",
    "    (humans4.Weight > 99.713)  |\n",
    "    (humans4.Weight < 46.083))\n",
    "humans4[cond] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Multivariate Outliers "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>BMI</th>\n",
       "      <th>zscore_BMI</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>167.089607</td>\n",
       "      <td>64.806216</td>\n",
       "      <td>23.212279</td>\n",
       "      <td>-0.620410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>181.648633</td>\n",
       "      <td>78.281527</td>\n",
       "      <td>23.724388</td>\n",
       "      <td>-0.359761</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>176.272800</td>\n",
       "      <td>87.767722</td>\n",
       "      <td>28.246473</td>\n",
       "      <td>1.941852</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>173.270164</td>\n",
       "      <td>81.635672</td>\n",
       "      <td>27.191452</td>\n",
       "      <td>1.404877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24996</th>\n",
       "      <td>163.952580</td>\n",
       "      <td>68.936137</td>\n",
       "      <td>25.645456</td>\n",
       "      <td>0.618008</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24997</th>\n",
       "      <td>164.334317</td>\n",
       "      <td>67.830516</td>\n",
       "      <td>25.117048</td>\n",
       "      <td>0.349063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24998</th>\n",
       "      <td>171.524117</td>\n",
       "      <td>75.861686</td>\n",
       "      <td>25.785295</td>\n",
       "      <td>0.689182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24999</th>\n",
       "      <td>174.949129</td>\n",
       "      <td>71.620899</td>\n",
       "      <td>23.400018</td>\n",
       "      <td>-0.524856</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>25000 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight        BMI  zscore_BMI\n",
       "0      167.089607  64.806216  23.212279   -0.620410\n",
       "1      181.648633  78.281527  23.724388   -0.359761\n",
       "2      176.272800  87.767722  28.246473    1.941852\n",
       "3      173.270164  81.635672  27.191452    1.404877\n",
       "...           ...        ...        ...         ...\n",
       "24996  163.952580  68.936137  25.645456    0.618008\n",
       "24997  164.334317  67.830516  25.117048    0.349063\n",
       "24998  171.524117  75.861686  25.785295    0.689182\n",
       "24999  174.949129  71.620899  23.400018   -0.524856\n",
       "\n",
       "[25000 rows x 4 columns]"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans5 = humans4[['Height', 'Weight']].copy()\n",
    "# Convert weight from cm to m\n",
    "humans5['BMI'] = humans5.Weight / (humans5.Height/100)**2\n",
    "humans5[\"zscore_BMI\"] = (\n",
    "    (humans5.BMI - humans5.BMI.mean()) / \n",
    "     humans5.BMI.std()\n",
    ")\n",
    "humans5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "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>Height</th>\n",
       "      <th>Weight</th>\n",
       "      <th>BMI</th>\n",
       "      <th>zscore_BMI</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>21388</th>\n",
       "      <td>165.912597</td>\n",
       "      <td>90.579409</td>\n",
       "      <td>32.905672</td>\n",
       "      <td>4.313253</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23456</th>\n",
       "      <td>187.110000</td>\n",
       "      <td>52.920000</td>\n",
       "      <td>15.115616</td>\n",
       "      <td>-4.741383</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23457</th>\n",
       "      <td>158.330000</td>\n",
       "      <td>92.780000</td>\n",
       "      <td>37.010755</td>\n",
       "      <td>6.402625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24610</th>\n",
       "      <td>169.082822</td>\n",
       "      <td>47.250297</td>\n",
       "      <td>16.527439</td>\n",
       "      <td>-4.022805</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Height     Weight        BMI  zscore_BMI\n",
       "21388  165.912597  90.579409  32.905672    4.313253\n",
       "23456  187.110000  52.920000  15.115616   -4.741383\n",
       "23457  158.330000  92.780000  37.010755    6.402625\n",
       "24610  169.082822  47.250297  16.527439   -4.022805"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "humans5[humans5.zscore_BMI.abs() > 4]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exercises"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### A Famous Experiment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "`summarise()` ungrouping output (override with `.groups` argument)\n",
      "\u001b[90m# A tibble: 5 x 3\u001b[39m\n",
      "   Expt  Mean Count\n",
      "  \u001b[3m\u001b[90m<int>\u001b[39m\u001b[23m \u001b[3m\u001b[90m<dbl>\u001b[39m\u001b[23m \u001b[3m\u001b[90m<int>\u001b[39m\u001b[23m\n",
      "\u001b[90m1\u001b[39m     1  909     20\n",
      "\u001b[90m2\u001b[39m     2  856     20\n",
      "\u001b[90m3\u001b[39m     3  845     20\n",
      "\u001b[90m4\u001b[39m     4  820.    20\n",
      "\u001b[90m5\u001b[39m     5  832.    20\n"
     ]
    }
   ],
   "source": [
    "%%R -o morley\n",
    "data(morley)\n",
    "morley %>%\n",
    "    group_by(`Expt`) %>%\n",
    "    summarize(Mean = mean(Speed), Count = max(Run))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Misspelled Words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Levenshtein distance: 1\n",
      "Similarity ratio: 0.8\n"
     ]
    }
   ],
   "source": [
    "%%bash \n",
    "similarity David Davin"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Levenshtein distance: 3\n",
      "Similarity ratio: 0.4\n"
     ]
    }
   ],
   "source": [
    "%%bash\n",
    "similarity David Maven"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Levenshtein distance: 5\n",
      "Similarity ratio: 0.814814814815\n"
     ]
    }
   ],
   "source": [
    "%%bash\n",
    "similarity \"the quick brown fox jumped\" \\\n",
    "           \"thee quikc brown fax jumbed\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "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>Name</th>\n",
       "      <th>Height</th>\n",
       "      <th>Weight</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>James</td>\n",
       "      <td>167.089607</td>\n",
       "      <td>64.806216</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>David</td>\n",
       "      <td>181.648633</td>\n",
       "      <td>78.281527</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Barbara</td>\n",
       "      <td>176.272800</td>\n",
       "      <td>87.767722</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>John</td>\n",
       "      <td>173.270164</td>\n",
       "      <td>81.635672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Michael</td>\n",
       "      <td>172.181037</td>\n",
       "      <td>82.760794</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Name      Height     Weight\n",
       "0    James  167.089607  64.806216\n",
       "1    David  181.648633  78.281527\n",
       "2  Barbara  176.272800  87.767722\n",
       "3     John  173.270164  81.635672\n",
       "4  Michael  172.181037  82.760794"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "names = pd.read_csv('data/humans-names.csv')\n",
    "names.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Elizabeth    1581\n",
       "Barbara      1568\n",
       "Jessica      1547\n",
       "Jennifer     1534\n",
       "             ... \n",
       "Josep           1\n",
       "iWlliam         1\n",
       "Joseeph         1\n",
       "eJennifer       1\n",
       "Name: Name, Length: 249, dtype: int64"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "names.Name.value_counts()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}