ch_11/1-EDA_unlabeled_data.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exploratory data analysis on unlabeled data\n",
    "We don't have labeled data yet, but we can still examine the data to see if there is something that stands out. \n",
    "\n",
    "## Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For this chapter, we will be reading data from a SQLite database:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>source_ip</th>\n",
       "      <th>username</th>\n",
       "      <th>success</th>\n",
       "      <th>failure_reason</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>datetime</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018-01-01 00:05:32.988414</th>\n",
       "      <td>223.178.55.3</td>\n",
       "      <td>djones</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-01 00:08:00.343636</th>\n",
       "      <td>223.178.55.3</td>\n",
       "      <td>djones</td>\n",
       "      <td>0</td>\n",
       "      <td>error_wrong_password</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-01 00:08:01.343636</th>\n",
       "      <td>223.178.55.3</td>\n",
       "      <td>djones</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-01 01:06:59.640823</th>\n",
       "      <td>208.101.11.88</td>\n",
       "      <td>wbrown</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-01 02:40:47.769630</th>\n",
       "      <td>11.76.99.35</td>\n",
       "      <td>tkim</td>\n",
       "      <td>1</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                source_ip username  success  \\\n",
       "datetime                                                      \n",
       "2018-01-01 00:05:32.988414   223.178.55.3   djones        1   \n",
       "2018-01-01 00:08:00.343636   223.178.55.3   djones        0   \n",
       "2018-01-01 00:08:01.343636   223.178.55.3   djones        1   \n",
       "2018-01-01 01:06:59.640823  208.101.11.88   wbrown        1   \n",
       "2018-01-01 02:40:47.769630    11.76.99.35     tkim        1   \n",
       "\n",
       "                                  failure_reason  \n",
       "datetime                                          \n",
       "2018-01-01 00:05:32.988414                  None  \n",
       "2018-01-01 00:08:00.343636  error_wrong_password  \n",
       "2018-01-01 00:08:01.343636                  None  \n",
       "2018-01-01 01:06:59.640823                  None  \n",
       "2018-01-01 02:40:47.769630                  None  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import sqlite3\n",
    "\n",
    "with sqlite3.connect('logs/logs.db') as conn:\n",
    "    logs_2018 = pd.read_sql(\n",
    "        \"\"\"\n",
    "        SELECT * \n",
    "        FROM logs \n",
    "        WHERE datetime BETWEEN \"2018-01-01\" AND \"2019-01-01\";\n",
    "        \"\"\", \n",
    "        conn, parse_dates=['datetime'], index_col='datetime'\n",
    "    )\n",
    "logs_2018.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## EDA\n",
    "The `success` column is now an integer because SQLite doesn't have Booleans:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "source_ip         object\n",
       "username          object\n",
       "success            int64\n",
       "failure_reason    object\n",
       "dtype: object"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logs_2018.dtypes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We are working with data for all of 2018, so it's important to keep an eye on memory usage:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "DatetimeIndex: 38700 entries, 2018-01-01 00:05:32.988414 to 2018-12-31 23:29:42.482166\n",
      "Data columns (total 4 columns):\n",
      " #   Column          Non-Null Count  Dtype \n",
      "---  ------          --------------  ----- \n",
      " 0   source_ip       38700 non-null  object\n",
      " 1   username        38700 non-null  object\n",
      " 2   success         38700 non-null  int64 \n",
      " 3   failure_reason  11368 non-null  object\n",
      "dtypes: int64(1), object(3)\n",
      "memory usage: 1.5+ MB\n"
     ]
    }
   ],
   "source": [
    "logs_2018.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The most common failure reason is providing the wrong password. We can also see that the unique usernames tried is well over the amount of users in our user base (133), indicating some suspicious activity:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>source_ip</th>\n",
       "      <th>username</th>\n",
       "      <th>success</th>\n",
       "      <th>failure_reason</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>38700</td>\n",
       "      <td>38700</td>\n",
       "      <td>38700.000000</td>\n",
       "      <td>11368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>4956</td>\n",
       "      <td>1797</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>168.123.156.81</td>\n",
       "      <td>wlopez</td>\n",
       "      <td>NaN</td>\n",
       "      <td>error_wrong_password</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>314</td>\n",
       "      <td>387</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.706253</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.455483</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             source_ip username       success        failure_reason\n",
       "count            38700    38700  38700.000000                 11368\n",
       "unique            4956     1797           NaN                     3\n",
       "top     168.123.156.81   wlopez           NaN  error_wrong_password\n",
       "freq               314      387           NaN                  6646\n",
       "mean               NaN      NaN      0.706253                   NaN\n",
       "std                NaN      NaN      0.455483                   NaN\n",
       "min                NaN      NaN      0.000000                   NaN\n",
       "25%                NaN      NaN      0.000000                   NaN\n",
       "50%                NaN      NaN      1.000000                   NaN\n",
       "75%                NaN      NaN      1.000000                   NaN\n",
       "max                NaN      NaN      1.000000                   NaN"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logs_2018.describe(include='all')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Distinct users per IP address\n",
    "Most IP addresses are associated with a single user, but at least one is associated with many:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    4956.000000\n",
       "mean        1.146287\n",
       "std         1.916782\n",
       "min         1.000000\n",
       "25%         1.000000\n",
       "50%         1.000000\n",
       "75%         1.000000\n",
       "max       129.000000\n",
       "Name: username, dtype: float64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logs_2018.groupby('source_ip').agg(dict(username='nunique'))\\\n",
    "    .username.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculate metrics per IP address\n",
    "The top 5 rows of the pivot seem to be valid users, since they have relatively high success rates:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>failure_reason</th>\n",
       "      <th>attempts</th>\n",
       "      <th>error_account_locked</th>\n",
       "      <th>error_wrong_password</th>\n",
       "      <th>error_wrong_username</th>\n",
       "      <th>success</th>\n",
       "      <th>success_rate</th>\n",
       "      <th>error_rate</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>source_ip</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>168.123.156.81</th>\n",
       "      <td>314</td>\n",
       "      <td>0</td>\n",
       "      <td>37</td>\n",
       "      <td>0</td>\n",
       "      <td>277</td>\n",
       "      <td>0.882166</td>\n",
       "      <td>0.117834</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24.112.17.125</th>\n",
       "      <td>309</td>\n",
       "      <td>0</td>\n",
       "      <td>37</td>\n",
       "      <td>0</td>\n",
       "      <td>272</td>\n",
       "      <td>0.880259</td>\n",
       "      <td>0.119741</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16.118.156.50</th>\n",
       "      <td>289</td>\n",
       "      <td>0</td>\n",
       "      <td>41</td>\n",
       "      <td>1</td>\n",
       "      <td>247</td>\n",
       "      <td>0.854671</td>\n",
       "      <td>0.145329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25.246.225.197</th>\n",
       "      <td>267</td>\n",
       "      <td>0</td>\n",
       "      <td>43</td>\n",
       "      <td>0</td>\n",
       "      <td>224</td>\n",
       "      <td>0.838951</td>\n",
       "      <td>0.161049</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30.67.241.95</th>\n",
       "      <td>265</td>\n",
       "      <td>0</td>\n",
       "      <td>37</td>\n",
       "      <td>0</td>\n",
       "      <td>228</td>\n",
       "      <td>0.860377</td>\n",
       "      <td>0.139623</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "failure_reason  attempts  error_account_locked  error_wrong_password  \\\n",
       "source_ip                                                              \n",
       "168.123.156.81       314                     0                    37   \n",
       "24.112.17.125        309                     0                    37   \n",
       "16.118.156.50        289                     0                    41   \n",
       "25.246.225.197       267                     0                    43   \n",
       "30.67.241.95         265                     0                    37   \n",
       "\n",
       "failure_reason  error_wrong_username  success  success_rate  error_rate  \n",
       "source_ip                                                                \n",
       "168.123.156.81                     0      277      0.882166    0.117834  \n",
       "24.112.17.125                      0      272      0.880259    0.119741  \n",
       "16.118.156.50                      1      247      0.854671    0.145329  \n",
       "25.246.225.197                     0      224      0.838951    0.161049  \n",
       "30.67.241.95                       0      228      0.860377    0.139623  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pivot = logs_2018.pivot_table(\n",
    "    values='success', index='source_ip', \n",
    "    columns=logs_2018.failure_reason.fillna('success'), \n",
    "    aggfunc='count', fill_value=0\n",
    ")\n",
    "pivot.insert(0, 'attempts', pivot.sum(axis=1))\n",
    "pivot = pivot.sort_values('attempts', ascending=False).assign(\n",
    "    success_rate=lambda x: x.success / x.attempts,\n",
    "    error_rate=lambda x: 1 - x.success_rate\n",
    ")\n",
    "pivot.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visual Anomaly Detection\n",
    "Let's see if something jumps out at us when plotting successes versus attempts by IP address:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:title={'center':'successes vs. attempts by IP address'}, xlabel='attempts', ylabel='success'>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAA5PElEQVR4nO3deXxcd3no/89zzuwarZYsy6vixHHiJGTBgUADpEmBkMJNaWka2gIF2rT3Qpd7KZcAtyXthRb6a+HSjRIulKVQtkLJZacBSkNjcALO6iyOo9iWZVnWOtKs55zn98c5MxkrkiM7kmYkPe/XSy+dZZbvOWfmPPPdRVUxxhhjAJxGJ8AYY0zzsKBgjDGmxoKCMcaYGgsKxhhjaiwoGGOMqbGgYIwxpsaCgjGrlIhcJSJHGp2O0yEivyEid5xi//dF5DeXM01rjQUFY+qIyC0i8k+ztjXsRtQsN3YR+biIvDta7hcRFZHp6G9ARG5udBrN4og1OgHGmBWrQ1U9EXkecLuI7FPVbzYqMSISU1WvUe+/WlhOoYmJyNtEZFBEciLysIhcE22v/WqL1k/6NSkiW0TkSyIyIiKjIvK3dft+S0T2R6/5oIhcFm3fKCL/Ej3ncRH5vbrnPEdE7hKRKREZFpH3R9tTIvJP0XtMiMheEemN9rWLyEdFZCg6hneLiBvtO0dE/l1EJkXkhIh8bp7j/4aIvHnWtntE5Bcl9AEROR6l6z4RuXCB5/WDInI4et7dIvKCaPu1wDuAX4l+Ad8jIu8BXgD8bbTtb6PHnici3xGRseja3FD3+h8Xkb+P0j8tIj8UkQ0i8n9EZFxEHhKRS+sePyAib4+ux7iI/GN0bluAbwAb636Vb5zvepzieN8RnecBEfm1aNvl0XPdusf9oojcs5BzWE9V7wQeAOY8/yLyBRE5Fl3vH4jIBXX71onIbdGx/Bg4e9ZzXxydr8no3Evdvt+Izu0HRGQUuEVEkiLylyJyKDq+fxCRdPT4bhH5avRZHROR/xARJ9o353dtTVJV+2vCP2AncBjYGK33A2dHyx8H3l332KuAI9GyC9wDfABoAVLAldG+XwYGgcsJv1znANsIfxzcDfwxkAC2AweBl0bPuxN4TbScBa6Iln8b+H9AJnrfZwNt0b4vAx+O0rAe+DHw29G+fwbeGb1vLX1znIPXAj+sW98FTABJ4KVRmjuiYzkf6Fvguf11YB1hTvktwDEgFe27BfinWY//PvCbdest0bV5ffQalwIngF111+dEdD5SwHeBx6PjcYF3A9+re70B4H5gC9AF/LB6feuvbd3j57wecxznVYAHvD86Zy8CZoCd0f4HgZfVPf7LwFvmea2P16WpH9Do2AX4GSAPXDPPc98AtEZp+D/Avrp9nwU+H53TCwk/n3dE+7qBHPAqIA789+h4fjPa/xvR+u9GaUkTfu5vi85jK+Hn88+jx/858A/Ra8UJg71wiu/aWvxreALsb54LE96wjwM/B8Rn7at9QaP12o0DeB4wAsTmeM1vAb8/x/bnAodmbXs78I/R8g+APwG6Zz3mDcB/As+atb0XKAHpum2vJroRAp8EbgU2P805aI1uYtui9fcAH4uWrwYeAa4AnGd4rseBi6PlW3j6oPArwH/MesyHgXfVXZ+P1O37XWB/3fpFwETd+gDwO3Xr1wGPzb62dfvnvB5zHNdV0U2zpW7b54E/ipbfBnw6Wu4ivLHPGViZOyhMROduP/B7CzzXHdFz2wkDZAU4r27/n/FkUHgtsKdunwBHODkoHJq1f4a6Gzrh9+HxaPlPga8A5yz0u7YW/6z4qEmp6gHgDwhvUsdF5LMisnEBT90CPKFzl61uAR6bY/s2wiKKieofYTFKb7T/jcC5wENREdHLo+2fIgw0nxWRoyLyFyISj14vDgzVvd6HCXMMAP+T8Av8YxF5QETeMM85yAFfA26MNr0a+HS077vA3wJ/R3h+bhWRtqc7OQAi8ocSFqFNRmlrJ/xVulDbgOfOOl+/Bmyoe8xw3XJhjvXsrNc8XLf8BHCqaz3f9ZjLuKrOzPPa/wS8IiqmuoEw0A2d4rVm61bVTlU9X1X/eq4HiIgrIu8VkcdEZIowAEJ4vnsIf+HPPvaqjfX7NLyD1z+WWes9hLnWu+uuyzej7QD/H3AA+LaIHJSocvwZfNdWJQsKTUxVP6OqVxLehBR4X7RrhvDDX1V/MzoMbBWRuRoRHGZWmW3d9sdVtaPur1VVr4vS8aiqvprwpv4+4Isi0qKqFVX9E1XdBTwfeDnhr7vDhDmF7rrXa1PVC6LXO6aqv6WqGwmLoP5eRM6Z5zT8M/BqCSszU8D36s7PX6vqswmLlc4F3jrPa9RE9Qf/k/Am2KmqHcAkT5ZVzzVs8Oxth4F/n3W+sqr6X5/u/U9hS93yVuDofOmZ73rM87qds/bVXltVBwmLon4ReA1hkF9svwpcT/grvJ0wlwHh+R4hzMnMPvaqofp9IiKzHgsnn58ThAH3grrr0q6qWQh/ZKjqW1R1O/BfgP9RrTs4xXdtzbGg0KREZKeIXC0iSaBI+GEPot37gOtEpEtENhD+yqn6MeGX6b0i0hJVWP5MtO//An8oIs+W0Dkisi16Ti6qbEtHv+4uFJHLo7T8uoj0qGpAWGQAEIjIz4rIRVFl5RRhUUAQ/dr8NvBXItImIo6InC0iL4pe75dFZHP0OuOEX8Lqsc32dcIv6p8Cn4vSUK0ofW6UM5mJztF8r1GvlfBGNALEROSPgfocxjDQX62ArNu2vW79q8C5IvIaEYlHf5eLyPkLeP/5vElENotIF2F9S7XyfRhYJyLt1QfOdz1O8dp/IiKJKCC+HPhC3b5PEgbJi4AvPYP0z6eV8AfCKOEPmT+r7lBVP3rPW0QkIyK7gNfVPfdrwAUSVoDHgN/j5B9AJ4nOx0eAD4jIegAR2SQiL42WXx595oXwh4BP+Dk+1XdtzbGg0LySwHsJf/0cI/xV+PZo36cIK5MHCG++tdY70RftFYTlpIcIy2B/Jdr3BcJy+c8QVuD9K9AVPeflwCWEFaInCANI9UZ0LfCAiEwDHwRuVNUC4Rf0i4QBYT/w7zz5a/O1hJXWDxLe+L8I9EX7Lgd+FL3ebYT1HAfnOgmqWiK8cfxclO6qNsIbwDhhkcMoYfFAtbXNN+Z6PcLirm8S1kc8QXgTqC+CqN4wR0XkJ9HyB4FXSdgy6K+jYq2XEBZrHSW8Pu8jvGZn6jOE1/IgYRHfuwFU9SHC3NLBqEhkI/Nfj7kcIzxHRwmL3n4nes2qLxMG3S+rav4ZpH8+nyQ8z4OEn4U9s/a/mbAo7RhhvcU/Vneo6gnCxhHvJby+Owgr4U/lbYRFRHui4qp/I6xIJnr+vwHThDmkv1fV73Hq79qaI2ExnTGmUURkgLDy9N8a9P6PEbYMa8j7m+ZiOQVj1jAR+SXC4rvvNjotpjlYj2Zj1igR+T5hJf1rqnU1xljxkTHGmBorPjLGGFOzoouPuru7tb+/v9HJMMaYFeXuu+8+oao9c+1b0UGhv7+fu+66q9HJMMaYFUVEnphvnxUfGWOMqbGgYIwxpsaCgjHGmBoLCsYYY2osKBhjjKmxoGCMMStMxQ+YLnlU/MXviL6im6QaY8xaM5IrsndgDD9QXEe4vL+LntbUor2+5RSMMWaFqPgBewfGSMdjdGdTpOMx9g6MLWqOwYKCMcasECUvwA+UVNwFIBV38QKl5FlQMMaYNaG+/iAZc3AdoVjxAShWfGKOkIwt3q3c6hSMMaZJzVV/cHl/F3sHxsiVKsSibXHXgoIxxqxq9fUHqbhLseKzd2CMF+/awIt3baDkhTmHxQwIYMVHxhjTlE5VfxB3HbLJ2KIHBLCgYIwxTWk56g/mYkHBGGOaUNx1uLy/i0LFY2S6SKHiLXr9wVysTsEYY5pUT2tqSesP5mJBwRhjmljcXZ5gUGXFR8YYY2osKBhjjKmxoGCMMabGgoIxxpgaCwrGGGNqLCgYY4ypsaBgjDGmZsmCgohsEZHviciDIvKAiPx+tP0WERkUkX3R33V1z3m7iBwQkYdF5KVLlTZjjDFzW8rOax7wFlX9iYi0AneLyHeifR9Q1b+sf7CI7AJuBC4ANgL/JiLnqqq/hGk0xhhTZ8lyCqo6pKo/iZZzwH5g0ymecj3wWVUtqerjwAHgOUuVPmOMMU+1LHUKItIPXAr8KNr0ZhG5V0Q+JiKd0bZNwOG6px3h1EHEGGOaTv1MaSvRkgcFEckC/wL8gapOAR8CzgYuAYaAvzrN17tJRO4SkbtGRkYWO7nGGHPGRnJFvvPgMb730DDfefAYI7lio5N02pY0KIhInDAgfFpVvwSgqsOq6qtqAHyEJ4uIBoEtdU/fHG07iareqqq7VXV3T0/PUibfGGMWrH6mtO5sinQ8xt6BsRWXY1jK1kcCfBTYr6rvr9veV/ewVwL3R8u3ATeKSFJEzgJ2AD9eqvQZY8xiOtVMaSvJUrY++hngNcB9IrIv2vYO4NUicgmgwADw2wCq+oCIfB54kLDl0pus5ZExZqWonymtOqfycsyUtthEVRudhjO2e/duveuuuxqdDGOMAcI6hb0DY3iBEnOEy/u76GlNNTpZTyEid6vq7rn22SQ7xhizSBoxU9pis6BgjDGLaLlnSltsKzflxhhjFp0FBWOMMTUWFIwxxtRYUDDGGFNjQcEYsyqs9DGHmoW1PjLGrHjV/gF+oLgL6B9Q8YMV3Wx0KVlQMMasaPVjDlV7Eu8dGOPFuzbMecM/3QCy1liINMasSNXioumSN+eYQ9V99cVJq2XQuqVkOQVjzIpT/2vfC5TJfJlU3CWbjFOs+OTLHnc8Gg6tX58bmD1oXdwVRqZ9pksenZlEIw+paVhQMMasCNV6AEeo/doveT77hyYYn6lwaLTAtnUZMskYvh+QTcaJu8JU0WPPwTGuvXDDSYPWlTyfe49MUKoEpOMOV2xfZ8VIWFAwxqwAJ+UMfCVfrtC2Ls49R6ZoTcZxHYcNrQkeOT7Nub1tHBzJE3MdhiaLBIGSK3pctKmN/u4sl/d3sefgGD95Yox0Isbu/i5ScfeU9RBrydo+emNM05tdD9CaivH4iQLj+QpBEI7yLMCxXIlUPMb61iRx1+G7Dx0nJg7peIx0IsZ9g5NU/ICe1hRX7uhm18Y2nnf2OjoyiRU798FSsKBgjGlqs+sBWpJx+rszFCs+uaJHruRxVncLFS8gHXdJJVx2bmilVAnIlTxKXsBFm9rR6LUAsskYLckYnh8GlZU698FSsOIjY0xTm2vymq6WBFftXM/l/Z3cNziJ5yteAP3rMsQch9ZUjLN7sly4qZ2OTBzPV7SitZt+3HW4vL+LvQNj5EqV2twHa73oCGySHWNMk6v4AYPjee4bnEThKZPXVCugc4UyPz08UZvgZnt3CwdPzJxywpu12onNJtkxxqxI9RXMIDxrUxubOjMn3cCr8xdkkzG6o2an1Zv82etbT3nTX+lzHywFCwrGmKY0V0/lB4em2NSZmfc5s2/ydtM/fXa2jDFNo35Qu9kVzNZCaHlYTsEY0xRmj0l06ZaOp1QwWwuhpWdn1xizrOYa4nquMYl+eniCS7d0MF2qcHh8hulSxVoILQPLKRhjls18I5ROlzxmSh6tqTgQFhVN5MuM58v4AQhC2EXNLDULucaYZTHfCKVDE3nuePQE+4dy3PnYKBP5MkfHZ/jJoXE++6PDPDKcI5uMkU3aiKbLwYKCMWZZzFVxXKoE3HlwlGwyxnPO6gKU7+4/xr/uO0qp4jMyXUJUefhYjpgrVtG8DCwoGGOWhSPV+oQKEA4tEW4XUnGX9nSC83pbGZwsIhIOiR1owJGJImU/IFf0rKJ5GVidgjFmyczubTyVL7N3YIyzu7OsyyZ59rYO7jkyyXi+TCIm/OTQOKmYS3smhiNC3HWZLlXIFWL4QcAV29dZRfMSW7KgICJbgE8CvYACt6rqB0WkC/gc0A8MADeo6riICPBB4DogD/yGqv5kqdJnjFla1UrlsqfcPzhJZ4vL/UdyjM+UuPuJMa67oI+S75MrVDh4YobBsQJlL6AlFeYacsUK+YrPpvYMr3v+VrZ1Zy0gLIOlPMMe8BZV3QVcAbxJRHYBNwO3q+oO4PZoHeBlwI7o7ybgQ0uYNmPMEsqXPf7j0RHirkNbOo6ifOO+YYYm8xwaKzI+7fHN+49x36EJxmbKdGUS7NrYRmdLgv51WUZyRTpSCbZ2Zrjphf2c09tmAWGZLFlOQVWHgKFoOSci+4FNwPXAVdHDPgF8H3hbtP2TGo7Qt0dEOkSkL3odY8wKMZIrcsejJ7j3yCQd6QTnrG8BhYmZMplkHNdVEjGXfNnnyHieHcl2Yo7S05qi4ivJuENXJsk5vVmu2tlDX8f8w1qYxbcsdQoi0g9cCvwI6K270R8jLF6CMGAcrnvakWjbSUFBRG4izEmwdevWpUu0Mea0VXMImUSMjkwcUThwfIbNXSkUoeL5CA7JmIOnSiCgGhCPxZguhv0UzutrpVD2eemFG8gkrNpzuS15fkxEssC/AH+gqlP1+6JcwWmN3a2qt6rqblXd3dPTs4gpNcY8E0MTeb7y00H2HZrgoaEcfe0pVJTxQplsMsH1l2wgk4zRmopRCZSebILNHRm297TQ25ZiqlihNeVS8QOu3NFtAaFBlvSsi0icMCB8WlW/FG0erhYLiUgfcDzaPghsqXv65mibMaaB6uccAOYcinpoIs9H7xjAERidKZGKOQxNwHl9bbVf/TOlHvraM9x/dJK467BjfSsvuaCX7tYUg+N5ssl41DHNei430lK2PhLgo8B+VX1/3a7bgNcB743+f6Vu+5tF5LPAc4FJq08wprHqh6WYKXugSksyftIQFRU/YM/BMWIO9LSmiLsOA6PTdLUE5MtpXrCjh0wiRiYR48bnbmO65AHhlJhx16HiBzw4NEV7Ol4b+G7vwBgv3rXBKpcbYClzCj8DvAa4T0T2RdveQRgMPi8ibwSeAG6I9n2dsDnqAcImqa9fwrQZY55G/bAUcVfYPzQFCM87uwXP19qNu+QFiAiphEvZC+hqSVD2M2zvbuHaC/tOKgaKuw6dmcRJ7zNXT+dcqULJCywoNMBStj66g/nzgdfM8XgF3rRU6THGnJ76m3UhGrY6rCxW0onwxj1d8vD8gIrvs6Uzw+HxPBOFMqpw1c6eBdULzDUHs/VcbhyryTHGzKn+Zp2Ixh0CcBxlPF8mV6jw9fuO8tjxGQplHwF2bmhlW1ecK7Z3Lbgpadx1uLy/i70DY+RKldp8ypZLaAwLCsaYOZ18s1a2rsuQK1T494dHUIRS2ScRFza0pQGYKpZpScbOqClpT2uqVhQ133zKZnlYUDDGzKv+Zu0I3L5/mE2dLYjAfx44wbGpIps6MriOEHNdApTgtBqZP8nmU24OdgWMMXOqzpAGYUuh6s3eEdh/dIrjuRJDE0VGpoqUo/qHdNy1uoAVznIKxpinGMkV2XNwtFb5e8X2dXRkEoBw75EJWpNxzunJUvI8Hjme46xAOa8va6OYrgIWFIwxNdXcwbcfGGJwooTrSNhHoVThl569lYs2tbHv0Diu4xN3HV512RZyJY8X7Oihty1lAWEVsKBgjAGe7Kg2Vahw+/4Rzt/QSkc6QaHi88DRHFefX2ZTZ4aLNrcTKHS2xEEFxxELCKuIBQVjzKyOag6uA0cmiriuw+MnphnPl/nu/mEu2dJB0fN55Ng0AOf2tnDN+dbzeDWxoGDMGlM/llH1Zl7fUc0PhP6eFg4ez7Pv8BjTpYCOdIyDI9Pce3iSnz1/PVefl2Gq6OEHGtU1mNXCgoIxa0j9WEb14xfN7lV8QV87+ZLP4yMemzrSnNPTgiI8MjyNKrhOOFzFyHTRhqNYZexKGrNG1BcRdWdTpOMx9g6MUfGDWke1QsVjZLpIIubw61ds4+yeLDt7W8mm4sRcQUQoeT6ADUexSllOwZg1Yq6B5yby5XA6zJbEU3oVA1ywuY1Do3lKvo8fKM8/uwuAkemiDUexSllQMGaNmF1ENDxV4IGjU4hAIubUipLqb/LXnNfLnoNjFCo+6bjLFdu76MgkbDiKVcyCgjFrRP1YRhP5Mg8cneLCjW30tqXnncOgpzXFtRc+dUwiCwarlwUFY1ahuVoYwZNjGY3NlBGB3mgwu1PNYWBjEq0tFhSMWWXma2FUFXcduloSJGKOzWFgnsI+AcasIqdqYVRvdmujQsWzSmMDWE7BmFXldKa2tDkMzFwsKBizwuTLHrmiR2sq9pTJbE53akurLzCzWVAwZgV5aGiSL949iBcExByHVz17E+f1tdf2x12HS7d0cOfBUUTCYGDFQuZ02CfFmBUiX/b44t2DtCZjbOnM0JqM8YW7BjmeK9bqDEZyRX56eAJHHFC4dEvHSZXMxjwdyykYs0Lkih5eEJBNRV9bgUePT/Gt+4dYl01y6ZYOfnp4gnQ8RmcmLDr66eEJulttWGuzcAv6pIjIL4tIa7T8v0TkSyJy2dImzRhTrzUVI+Y4TBc9fFUeOTZFOhFjc1eGdDzGnoNjlL3gpEpmL1BKXvA0r2zMkxb68+GPVDUnIlcCPwd8FPjQ0iXLGANPzoRW8QMyiRivevYmciWPx0/MMFP2ef5Z64g5Ti0QBKoUKzZgnTlzCy0+8qP/Pw/cqqpfE5F3L1GajFn15utxXL9/cDzPfYOTALVOaOf1tfM/XtLCgeEpvnT3UYZyRcYLFfrXZUjGnVoRUq5UsQHrzBlZaFAYFJEPAy8G3iciSayS2pgz8nQ9jkdyRe44cIK7nxgjm4xx8eZOUnH3pLGJjkwUuWRbBwOjMxTKHvcfneKNV/bT15GhuzVlfQ/MGVtoULgBuBb4S1WdEJE+4K1LlyxjVqf6HsfVfgT1N/uKH3D7Q8M8NjzDSK5MruDh++O8cGcPXunJ+gE/UHrb0nRnk5R9ZTJfpjUdzoBmfQ/MM7HQT04f8DVVfVRErgJ+GfjxqZ4gIh8TkeMicn/dtltEZFBE9kV/19Xte7uIHBCRh0Xkpad/KMY0v7l6HNdXBk+XPB45Nk17Jk426Ya5gvE8E/lyrX6gvoOa6zgIkIw7VndgFsVCP0X/Avgicg5wK7AF+MzTPOfjhLmL2T6gqpdEf18HEJFdwI3ABdFz/l5E3AWmzZgVo/6GDnNXBgvgirC1q4WK75Ov+BQrfq1+wMYtMktpocVHgap6IvKLwN+o6t+IyE9P9QRV/YGI9C/w9a8HPquqJeBxETkAPAe4c4HPN2ZFqJ/TYK7K4Gwyxo7eFg6NFXEdYWNHmsu2pvn5Z208aUgLG7fILJWFBoWKiLwaeC3wimhb/Azf880i8lrgLuAtqjoObAL21D3mSLTtKUTkJuAmgK1bt55hEoxpnI5Mgued3Q2EQaD+hh53Ha45fwN7Do7WzXa27iljHFUfa8HALLaFBoXXA78DvEdVHxeRs4BPncH7fQj434BG//8KeMPpvICq3kpYhMXu3bv1DNJgzLKYq9np07U8gupsZ32WCzANsaCgoKoPisjbgK3R+uPA+073zVR1uLosIh8BvhqtDhLWU1RtjrYZsyLNdfPvyCRO2fKonuUCTKMsdJiLVwD7gG9G65eIyG2n+2ZRU9aqVwLVlkm3ATeKSDLKhezgaVo3GdOs5pvoZrrknbLlkTHNYKHFR7cQVvx+H0BV94nI9lM9QUT+GbgK6BaRI8C7gKtE5BLC4qMB4Lej13tARD4PPAh4wJtU1Z/jZY1pahU/YGymTKHsk4rF8IKgNtENcFpzHRjTCAuuaFbVSRGp33bKnzeq+uo5Nn/0FI9/D/CeBabHmKZTLTI6kSvx/YdH2NyZpqslSf+6DIm4QzYZO2XLI2OawUKDwgMi8quAKyI7gN8D/nPpkmXMylEdtG7PwVHScZexfJmz12c5OpEnEXNqQ1DEXceakpqmt9Cg8LvAO4ESYae1bwE2IJ5Z86q5g5mSz0NDU1y4qZ0gUDa0pUi4Drs2tVOu+LUhKMAqkU1zW2jrozxhUHjn0ibHmJWjvkK5LRXn8RPTPDI8RcwN5zyIuw4JVxBsCAqzciy09dF3RKSjbr1TRL61ZKkypslVK5Srk9q4jsOzNndQrAR0pOPkSh7drXEqfmD1BmZFWWjxUbeqTlRXVHVcRNYvTZKMaV718xz4Adw/OAUKve1pkjGXy7Z1ceWObpIxh0CxegOz4ix47CMR2aqqhwBEZBths1Jj1oyRXJE9B8e4+4kxYo7DJVvauXBTG/cPTqESBoArtnfRmUk8/YsZ06QWGhTeCdwhIv9OOIjjC4jGHzJmLajWH8wUKxyfCger+8Z9BV52UR8XbGzj+Wd309WSsFyBWfEWWtH8TRG5DLgi2vQHqnpi6ZJlTHMpeQFlTzk6WSTuOmQSMWZKHvsOT3DxlnYLCGbVWGhF8ysJO7B9VVW/Cngi8gtLmjJjmkgy5qCqlD2f7T1Z8uUKFT+g4vlctKndAoJZNRb6SX6Xqk5WV6JK53ctSYqMaUJxN6wv8ALIl322rcvyop097D5rHZs6M41OnjGLZqF1CnMFj4U+15hVoa8jwxuv7OfOg6OIhGMWWXNTs9os9MZ+l4i8H/i7aP1NwN1LkyRjGq86dEXFD4i7Tm0ynL6ODK+4OGXDVJhV63SGufgj4HPR+ncIA4Mxq0q1H8IPHzvB/YOTDE+W6G1NcvHWdq45fwM9rSkbpsKsagttfTQD3LzEaTFmSc01E1q9kVyROw6c4K6BMY5NFIm50NWSIF8JePxEgT0HR7n2wj4LCGZVW1BQEJHvMUdnNVW9etFTZMwSeLppMCt+wO0PDfPY8AzDU0VGZ8oISnc2hR+EU3sUKj4lL7CgYFa1hRYf/WHdcgr4JcLJcIxpevUD1803DeZ0yeORY9N0tSRom4kzVfA4kSsxVazgSPiYdNy1ge3MqrfQ4qPZlco/FBGbLtOsCCUvOGkazLgrjEz7TJe8k4akEMAVYdu6FmbKFSYKZcZmSmzpbOGs7jRXbF9nuQSz6i20+KirbtUBdgPtS5IiYxZBff1BMubgOsJ0qcJMyWP/0BSer6TjDldsX0dPa4psMsaO3hYOjYVDWGzvzvLc/m6uvWgD6USs1vrImNVuocVHdxPWKQhQIZxf+Y1LlCZjnpG56g+2d7fwhbsGOXA8RyrhcvXOHrLJ+EnFSNecv4E9B0cpVHzScbcWMIxZSxYaFN4GfFNVp0Tkj4DLgPzSJcuYMzNX/cGeg2P4QcB5fVkcUdozCYYmS2zszOCVtFZ53NOa4toL+6wPglnTFvqp/19RQLgSuBr4v8CHli5ZxpyZav1B3BUKFZ+4KxyfKvCjg6M8MDjFwOgMk4UyZT8gV/SIOXJS5XF9RzVj1qKF5hT86P/PAx9R1a+JiM3RbJpOMuYwUw7rDWKOUPaUJ0ZzHBkvAsJkqczjI9Ns7mphU0eSl1xg/Q6MqbfQb8OgiHwY+BXg6yKSPI3nGrNsKn5AqexR9pRCOWAyX+bA8Rn6OlKUKj5Jx6ESwM+cs462dJIOmxDHmJMsNKdwA3At8JeqOiEifcBbly5Zxpy+kVyROx49wf1HpxieKhMEAa4jTMyUOZZwQQTXccg4Ab1tGRS1zmjGzLLQfgp54Et160PA0FIlypj5VJuaOsJJcyDnyx7/8egIpbLPgeMzTOTLuI6wqSNNRWEyXyYVd/F9SMZdHAccEeuMZswsNvy1WTGqTU3HZyocHJlmU0eKnrYUO9Zn2T+UY9+hSYZzRWKOUvKUiucxWaiQTToECJlkjEBhW1easufzgh09lkswZpYlCwoi8jHg5cBxVb0w2tZFONJqP2FfhxtUdVxEBPggcB1hU9ffUNWfLFXazMpTbWqacB0GRqc5Nlnk6ESBLV1p7jwwygt39tCainFsMmA875GJCwUcUnGXIFC292bZ1tnCxVvbCQJ46YUbyCTsN5Exsy3lz6SPE9ZD1LsZuF1VdwC38+TIqy8DdkR/N2HNXdes+nkMquvj+TLHp4oUyj6Fis/hsTwtyRgxVxAcjk4WcETY1deKSNh7ORWP4wiIwIb2FJs7UvgofqBcuaPbAoIx81iyb4aq/kBE+mdtvh64Klr+BPB9wo5x1wOfVFUF9ohIh4j0RXUXZo2Y3RN5e3cL+45M8NBQjlyhEuYM1mU4PFYgGSsRc4VAoewFlDyf7myKnzuvh6/e59PfleHoVIENbWk6WhJcsLGNQtm3HIIxT2O5vx29dTf6Y0BvtLwJOFz3uCPRNgsKa0S1eMgRQR1BVfnCXYNUfI+JQoXB8SIlz2dwPE+x4jFdVM7rawVV+tpT+IEyMl2kJR3nrS89l0eOT7O9mOXwWIHetgSBWg7BmIVo2DdEVVVEnjJHw9MRkZsIi5jYunXroqfLLK9qa6KKH/D4yAz3HhlHAV/B83xcR+hsSZFJuFR8n4ETBdJxISAc0bS/O0tfR4oX7VxP3HVqrZF29rXP2UrJGHNqyx0UhqvFQlFfh+PR9kFgS93jNkfbnkJVbwVuBdi9e/dpBxXTPOqLi05Ml/jKviO0JuP4qoxOl8iVPNIxYddGoeIrw1MlWpIu2VScoFDh8HiBjR1pYg5PGZrCpsw05sws97fmNuB10fLrgK/UbX+thK4AJq0+YXWrH7iuNRXn0eEcvkKu5DFwYoZcySPhOqgI+w5PUvI8XIHWZAwEurIJEKHsa1ibbIxZFEvZJPWfCSuVu0XkCPAu4L3A50XkjcAThD2lAb5O2Bz1AGGT1NcvVbpMc6gOXJcvVdh3eIKhiSKT+QrdmTjpeIxCxUOAy/u7OHB8hk2daVqTcTZ3ZshXPESEfNnnhTu7yZcD65lszCJZytZHr55n1zVzPFaBNy1VWkzzScYcxmfKfP+h46QSLiPTJRyBkZky06Vwpte2lEu5EtC/roUrtvdw7voWvvXAcY5M5EknYly9swdXXGKOWs9kYxaJNcUwy6Z+iIqj43n2HR5naKpIOu5S9nySMQcR6MrEGZ0pozgM58pcf/FGulriXLqtiws3dzBwYppHhqdxHKFQ8bi8v8tyCcYsEgsKZlmcPETFDMcm8zwxlkdQxqfLeEFAJuWigbCzr42pQpn1rUkScZf2lnjtxh93HXZt7GBHb5tNhmPMErCgYJZEfa5gpuTx/YePk0nEGBidplDxGBjLk3SFE3mPkhcwXfZRgc0daTZ2pOnvbjllhzNrXWTM0rCgYBZVxQ8YHM9z3+AkUwWPB4cmmSl6HB4vUPErjOQquK4wPlOmJ5tgOFcm5iiIS8oVvABmih5bNmatw5kxDWDfNrNoRnJF7jhwgrufGCPuOswUPZ4Ym2Z4ssx4vsxMySfuQCLuICqMTJdRVeIxl56WBB3ZBEnX4bdedBbrsikrGjKmASwomGesOojdN+4f4tFj0zx8LMexyQKTRY9SxccR8APwFAIfPD8gHhcclPZMDNdxaE0nCALo7UrSlk6QTdpH05hGsG+eeUaqFciD4wW+ePdhXIEjE0WCIOxsXvEJRyvlyZ6SmYQQj7skYy7betIUykpPNkmActnWTgsIxjSQffvMGasfxO7QWJ7pYgURCDyfigIaBQIF1wX1o87HEvZTOL+vla5sEgjXz9vQyksu6LMiI2MayIKCOWMlL2B8psSjwzP88MAII9NlPD/MFcQEsukYcVeYLHjEXYi5QjrmcM76Vq67eAPZZIKrz19PlKl4yvhFxpjlZ0HBnJZqU9NkzMERODCc5+HhHCO5IgRPBgQViDtKSyrBukycllScRNwFVTZ2ZujIJLli+zra04lGH5Ixpo4FBbNgQxN57jw4iqrgCGzqSHJsKs+DgxNMl5TqkLWBQm/G5YbL+8mmwgHvurNJVCDhOhzPlbhyRw+dGQsIxjQbCwpmQYYm8nz0jgGKnsdIrkRHOs7A6AzDEwWKXkD9GOYKTJV8TkyW6G5N8tjIDOm4S297mmLFJ5t0rTLZmCZl30zztE5MF7lt3yATM0UeH52hUPZ5sOiRL1Uoe1D2T368CyRiws5NrezobaMtHef+wSk0qmC2sYqMaV4WFMxT1A9RsefACB+/8wkGxwsMTRVQT/EI+x0E0eNdoBoXHKA1JaQTcbqyKQB629KowvPP7qarJWEBwZgmZkHBnGQkV2TPwTFGcgUOjsyw5+Ao2USMsucT+EoxeOpz6jMKAqTiLhdvbiMVD2/+xYpPMu5YQDBmBbCgYGoqfsC3HzzGY8MzHJ7IM5EvcyJXwsnCRN6rNR2dS1sCEokYLYkYv/qcfl60s5uDJ2YYmS4Sc8SKjIxZISwoGPJlj1zRY2h8hm/dfwyAw2N5KpWAQsVndKZExfepzMolCOAKZGKwa1MnF2xso7c9xS/v3kJnJsHZ61tteGtjVhgLCmvcQ0OTfOGuQUZniuwfylEoVShUAgrlCgGCKMyUfHyiIMCTxUUxgVQMNnW18MpLNtPVmqDiB7WWRTa8tTErjwWFNSxf9vjUnicYmy4znCsyXSwzPhP1Snag7CsCJGPQHndAhETMJRkTZkoBrakY527IcvGWDtIpl4ofWDGRMSucBYU17NCJaR4ZztGejjFZqDBZ8Ch4YT8D8cOcQMwNWxoVPWVLZxLHdelpTdHXnuKGZ29m58Z24q5jxUTGrBIWFNaYanPTvY+P8OEfDPDQ0BQVT3FEawHBIWxu6jiQijs4OKSTLol4nM6WBJva07z+yn7O6W2rva4FA2NWBwsKa0jY3HSUQ2MzfGbPIbIJl0Ch6D/ZrCgGpOJCyVdcR1jfmiKTcEFcrtrZzUTeY9fGNrZ1Zxt3IMaYJWNBYY2o+AHffmCIR4dneHh4inzZp+IHuKIk3LBXskOYO+jKhv0JNIDObIJULMamjiT5so/rCFees85yBsasUhYU1ojHjue47Z6jFMsBYzNlCiWPGQE0rDNwgGwCFAdXhC1dLbz5qu1kUnEeGZ5Go8zEFdu76OvINPJQjDFLyILCKlTtd+A64Q0/XyzzqT1PcHSiSCYRo+z7pJMxJgseEj1nXdoB1+Wc9Vm2dGZ4w5X9nNfXAcCO3jarSDZmjbCgsMo8NDTJ5/ce4dhUgYHRGc5a18JM2ac1FQ4zMVWoUCj7BAoJF1qSDl6gdLWm6W1Pcc156+nKpjh7/cmVyBYMjFkbLCisEhU/YCJf5mP/8Tij+TIDIzM4DhwZz9PdmuTgiQKOKvmyjyLEXehMJ1nfkaI16my2tauFrmySK7ZbXwNj1qqGBAURGQByhJ1jPVXdLSJdwOeAfmAAuEFVxxuRvpVmJFdk78AYj49M88ODJ2hPxchXPDozCaaKHhvaU/hBQGdLklzJJ+YISsDZvVkmChW2dGZwXYcbdm9mR2+bBQRj1rBG5hR+VlVP1K3fDNyuqu8VkZuj9bc1JmkrQ8UPmC557Dk4SjruMjRZpOT5jOeVUsVndKZMSzJGRyrOEcdhZ28b6YTL5s4Mg+MFtq5rIRkXztvQDmABwRjTVMVH1wNXRcufAL6PBYV5VYe4nixUOHA8x4Ub2xiZLrO5I8OxqRLpZIyJgkdvW5JMMsErnrWBretauay/kweOTtKVTRAESl97GtfBhqcwp6V+rm773KwujQoKCnxbRBT4sKreCvSq6lC0/xjQO9cTReQm4CaArVu3Lkdam07FD7j9oWEOjeYJFA6PzVAoewjK5o4M8ZhDWypG0Qv43Z89h2dt6WSm5LF3YAxflfP72rloUxvr21IEin2xzWmpFlf6QdjB8fL+LnpaU41OllkkjQoKV6rqoIisB74jIg/V71RVjQLGU0QB5FaA3bt3n2KE/9Wn+ussX/Z45Ng06YTLyFQBVxzuG5yivytJRYUL+tqJuQ5ndWd4dn/Y0SyTiPHiXRvs1515Rip+wN6BMdLxGKm4S7His3dgjBfv2mCfqVWiIUFBVQej/8dF5MvAc4BhEelT1SER6QOONyJtzar666zsKdPFClOFCqPTRTKJOF1ZB1+VizZ3sr4tha9KOu5yxfaTex5b01LzTJW8AD9QUnEXCGfZy5UqlLzAPlurxLIHBRFpARxVzUXLLwH+FLgNeB3w3uj/V5Y7bc2q+uus7AUMjM4wU/IZHM/jOkBW8Hylf12GDR1prtq5nrjrWG7ALIlkzMF1hGLFr+UUYo6QjNlnbbVoRE6hF/iyiFTf/zOq+k0R2Qt8XkTeCDwB3NCAtDWVanFRxQ8oe8rA6AxJ16W1LU6hJ8ujI9NM5Muk4i5eoJQ8n2wyZsHALJm463B5fxd7B8bIlSo21eoqtOxBQVUPAhfPsX0UuGa509Os6ivzQJjMl5kpeiRaHApln65ski2eTyYRJ5uKLqOuqSoW0yA9rSmrn1rFmqlJqonMVZl3dHyGR49P4zoOqZjDJVvbaU8nuPLcbjQQ4jFhPF+2sl2zLKx+avWyoNCEZlfmxV3h2FSJF5+/nsHJIhVfOTZZZntPBlRIJ6xs1xizOCwoNKHZlXlTRQ9HhK3dWbZ1Zyn7ymS+zKVbO3hwaMrKdo0xi8aCQhOaXZknCDt6W/D8MPcgvk8y7rCpM8OmzoyV7RpjFo0FhSY1uzJvIl+et8WHBQNjzGKxoNDE6ivzrMWHMWY5WFBYQazFhzFmqdkdxhhjTI0FBWOMMTUWFIwxxtRYUDDGGFNjQcEYY0yNBQVjjDE1FhSMMcbUWFAwxhhTY0HBGGNMjQUFY4wxNRYUjDHG1FhQWAUqfsB0yaPiB41OijFmhbMB8Va4+rmc3WhI7Z7WVKOTZYxZoSynsILVz+XcnU2RjsfYOzBmOQZjzBmzoLCCzZ7LORV38QKl5FlQMMacGQsKK1j9XM4AxYpPzBGSMbusxpgzY3ePBnqmFcTVuZwLFY+R6SKFinfSNJ3GGHO6rKK5QRargtim6TTGLCa7gzTAYlcQx12HbDJmAcEY84ytyZxC/81fO2n90fe8bFlvqHNVEOdKFUpeYDd2Y0xDNd0dSESuFZGHReSAiNy82K8/OyAA7HjnNxjJFRf7reZlFcTGmGbVVHchEXGBvwNeBuwCXi0iuxbr9ecKCFWXv+f2ZWvfbxXExphm1WzFR88BDqjqQQAR+SxwPfDgcrz5chbfWAWxMaYZNdudaBNwuG79SLStRkRuEpG7ROSukZGRRX3z5S6+sQpiY0yzWXF3I1W9VVV3q+runp6e03ruo+952Sn3283ZGLPWNdtdcBDYUre+Odq2KOKuw953XjPnvoH3/vxivY0xxqxYzRYU9gI7ROQsEUkANwK3LeYb9LSmnpJjsIBgjDGhpqpoVlVPRN4MfAtwgY+p6gOL/T5x17FAYIwxc2iqoACgql8Hvt7odBhjzFrUbMVHxhhjGsiCgjHGmBoLCsYYY2osKBhjjKkRVW10Gs6YiIwAT5zBU7uBE4ucnOVmx9B4Kz39YMfQDBqR/m2qOmfv3xUdFM6UiNylqrsbnY5nwo6h8VZ6+sGOoRk0W/qt+MgYY0yNBQVjjDE1azUo3NroBCwCO4bGW+npBzuGZtBU6V+TdQrGGGPmtlZzCsYYY+ZgQcEYY0zNmgsKInKtiDwsIgdE5OZGp2ehRGRARO4TkX0icle0rUtEviMij0b/OxudzioR+ZiIHBeR++u2zZleCf11dE3uFZHLGpfyJ81zDLeIyGB0HfaJyHV1+94eHcPDIvLSxqT6SSKyRUS+JyIPisgDIvL70fYVcx1OcQwr6TqkROTHInJPdAx/Em0/S0R+FKX1c9F0AYhIMlo/EO3vX9YEq+qa+SMcjvsxYDuQAO4BdjU6XQtM+wDQPWvbXwA3R8s3A+9rdDrr0vZC4DLg/qdLL3Ad8A1AgCuAHzU6/ac4hluAP5zjsbuiz1MSOCv6nLkNTn8fcFm03Ao8EqVzxVyHUxzDSroOAmSj5Tjwo+j8fh64Mdr+D8B/jZb/G/AP0fKNwOeWM71rLafwHOCAqh5U1TLwWeD6Bqfpmbge+ES0/AngFxqXlJOp6g+AsVmb50vv9cAnNbQH6BCRvmVJ6CnMcwzzuR74rKqWVPVx4ADh561hVHVIVX8SLeeA/YRznq+Y63CKY5hPM14HVdXpaDUe/SlwNfDFaPvs61C9Pl8ErhERWZ7Urr3io03A4br1I5z6A9ZMFPi2iNwtIjdF23pVdShaPgb0NiZpCzZfelfadXlzVLzysboiu6Y+hqgI4lLCX6kr8jrMOgZYQddBRFwR2QccB75DmIOZUFUvekh9OmvHEO2fBNYtV1rXWlBYya5U1cuAlwFvEpEX1u/UMK+5YtoXr7T01vkQcDZwCTAE/FVDU7MAIpIF/gX4A1Wdqt+3Uq7DHMewoq6DqvqqegnhvPPPAc5rbIrmt9aCwiCwpW59c7St6anqYPT/OPBlwg/WcDV7H/0/3rgULsh86V0x10VVh6MveAB8hCeLJpryGEQkTngz/bSqfinavKKuw1zHsNKuQ5WqTgDfA55HWDxXnf2yPp21Y4j2twOjy5XGtRYU9gI7olr/BGElzm0NTtPTEpEWEWmtLgMvAe4nTPvrooe9DvhKY1K4YPOl9zbgtVHrlyuAybrijaYyq4z9lYTXAcJjuDFqOXIWsAP48XKnr15UDv1RYL+qvr9u14q5DvMdwwq7Dj0i0hEtp4EXE9aNfA94VfSw2dehen1eBXw3ytEtj0bWyjfij7CFxSOEZXrvbHR6Fpjm7YQtKu4BHqimm7Cc8XbgUeDfgK5Gp7Uuzf9MmK2vEJaXvnG+9BK2zvi76JrcB+xudPpPcQyfitJ4L+GXt6/u8e+MjuFh4GVNkP4rCYuG7gX2RX/XraTrcIpjWEnX4VnAT6O03g/8cbR9O2HAOgB8AUhG21PR+oFo//blTK8Nc2GMMaZmrRUfGWOMOQULCsYYY2osKBhjjKmxoGCMMabGgoIxxpgaCwrGzENE3lG33CEi/20Z3rNfRH51qd/HmPlYUDBmfu+oW+4gHL1yqfUDFhRMw8Se/iHGrH4i8q+EQwukgA8SdixKR4OYPUA47PrZ0fp3VPWtIvJW4AbCYZq/rKrvigZt+yawB3g+YS/6fwT+BFgP/Jqq/lhEbiEcu+ccoBv4C1X9CPBe4PzofT4BfDt6foLwR9wvqeqjS3kuzNpmndeMIZx4RlXHomEI9gIvAp5Q1Wy0vx/4qqpeGK2/hHAIgt8m7Al8G+E8BYcIe6JeShhM9hL2RH8j8F+A16vqL0RB4ZWE4+q3EPZ4fS5wLuE8AS+P3udvgD2q+uloaBZXVQtLfDrMGmY5BWNCvycir4yWtxCOmXMqL4n+fhqtZ6PnHAIeV9X7AETkAeB2VVURuY+weKjqK9ENviAi3yMc1G1i1vvcCbxTRDYDX7JcgllqVqdg1jwRuQr4OeB5qnox4Y0+9XRPA/5cVS+J/s5R1Y9G+0p1jwvq1gNO/iE2O5v+lGy7qn6GMIdRAL4uIlc//REZc+YsKBgTDk08rqp5ETmPsEgHoBIN2wyQI5wOsupbwBuicf4RkU0isv403/f6aP7edcBVhEVNJ72PiGwHDqrqXxOOovms03wPY06LFR8ZE1YM/46I7CccWXNPtP1W4F4R+Ymq/pqI/FBE7ge+EVU0nw/cGc2UOA38OuCfxvveSzh8cjfwv1X1qIiMAL6I3AN8nLAS+zUiUiGcJe3PnunBGnMqVtFsTANEFc3TqvqXjU6LMfWs+MgYY0yN5RSMMcbUWE7BGGNMjQUFY4wxNRYUjDHG1FhQMMYYU2NBwRhjTM3/D0Y1jnJtWm2uAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pivot.plot(\n",
    "    kind='scatter', x='attempts', y='success', \n",
    "    title='successes vs. attempts by IP address', alpha=0.25\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can imagine there being a separation boundary between the groups by considering the fact that valid users probably have a close to 1:1 relationship between attempts and successes:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7de70b523fd0>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = pivot.plot(\n",
    "    kind='scatter', x='attempts', y='success', \n",
    "    title='successes vs. attempts by IP address', alpha=0.25\n",
    ")\n",
    "ax.plot([30, 350], [0, 340], 'r--', label='sample boundary')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "More IP addresses are attacker IP addresses because they get a new IP address for each attack, while the valid users stick with the 1-3 they have. This makes the outliers on successes the valid users instead of the attackers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "attempts       AxesSubplot(0.125,0.125;0.352273x0.755)\n",
       "success     AxesSubplot(0.547727,0.125;0.352273x0.755)\n",
       "dtype: object"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pivot[['attempts', 'success']].plot(\n",
    "    kind='box', subplots=True, figsize=(10, 3),\n",
    "    title='stats per IP address'\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Does looking at this on a minute resolution make anything stand out?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib.ticker import MultipleLocator\n",
    "\n",
    "ax = logs_2018.loc['2018-01'].assign(\n",
    "    failures=lambda x: 1 - x.success\n",
    ").groupby('source_ip').resample('1min').agg(\n",
    "    {'username': 'nunique', 'success': 'sum', 'failures': 'sum'}\n",
    ").assign(\n",
    "    attempts=lambda x: x.success + x.failures\n",
    ").dropna().query('attempts > 0').reset_index().plot(\n",
    "    y=['attempts', 'username', 'failures'], kind='hist',\n",
    "    subplots=True, layout=(1, 3), figsize=(20, 3),\n",
    "    title='January 2018 distributions of minutely stats by IP address'\n",
    ")\n",
    "for axes in ax.flatten():\n",
    "    axes.xaxis.set_major_locator(MultipleLocator(1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There seems to be something with the number of usernames with failures per minute that we can use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'usernames with failures')"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "logs_2018.loc['2018'].assign(\n",
    "    failures=lambda x: 1 - x.success\n",
    ").query('failures > 0').resample('1min').agg(\n",
    "    {'username': 'nunique', 'failures': 'sum'}\n",
    ").dropna().rename(\n",
    "    columns={'username': 'usernames_with_failures'}\n",
    ").usernames_with_failures.plot(\n",
    "    title='usernames with failures per minute in 2018',\n",
    "    figsize=(15, 3)\n",
    ").set_ylabel('usernames with failures')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<hr>\n",
    "<div style=\"overflow: hidden; margin-bottom: 10px;\">\n",
    "    <div style=\"float: left;\">\n",
    "        <a href=\"../../ch_10/red_wine.ipynb\">\n",
    "            <button>&#8592; Chapter 10</button>\n",
    "        </a>\n",
    "        <a href=\"./0-simulating_the_data.ipynb\">\n",
    "            <button>Simulation</button>\n",
    "        </a>\n",
    "    </div>\n",
    "    <div style=\"float: right;\">\n",
    "        <a href=\"./2-unsupervised_anomaly_detection.ipynb\">\n",
    "            <button>Next Notebook &#8594;</button>\n",
    "        </a>\n",
    "    </div>\n",
    "</div>\n",
    "<hr>"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}