Data_preparation_and_cleaning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d46160cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "class color:\n",
    "   PURPLE, CYAN, DARKCYAN, BLUE = '\\033[95m', '\\033[96m', '\\033[36m', '\\033[94m'\n",
    "   GREEN, YELLOW, RED, BOLD = '\\033[92m', '\\033[93m', '\\033[91m','\\033[1m'\n",
    "   UNDERLINE, END = '\\033[4m','\\033[0m'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "78599b80",
   "metadata": {},
   "source": [
    "## What was done in file_reduction.ipynb:\n",
    "Every touch events was tied to a second. Then, each row of the dataframe would represent how many times one user touched their phone in one second. \n",
    "Now, spikes in the data (defined as the statistical outliers) have been treated and moved to the median of the distribution. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a863e51f",
   "metadata": {},
   "source": [
    "# Cleaning Data: reasonable results and time-interpretable grouping.\n",
    "In order to get in second 40 touches per second you need to use all ten fingers on the screen, four times, or alternatively just your thumbs 20 times. In one second. Which is just impossible. Consider this: \n",
    "- the fastest finger tapping rate was set by Japanese musician Kaoru Watanabe in 2014. He achieved a rate of 13.36 taps per second.\n",
    "- Prabhshish Singh (at 13) accurately typed letters from A to Z on a mobile phone in just 3 seconds and 44 milliseconds, as confirmed on August 29, 2022, making it about 7.5 touches each second. \n",
    "\n",
    "Not all users here must be world-record breakers, so for now statistical outliers will be treated as data spikes derived from device malfunction or improper handling of the screen.\n",
    "\n",
    "We need to consider the **extreme touches**. If a value is an outlier it needs to be subsituted with the median. The median is a much more defensible measure because it has much less impact on the statistics than choosing any value, i.e. of the third quartile of the touch distribution. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "2647ba83",
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy import stats\n",
    "from tqdm import tqdm\n",
    "import os\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import plotly.express as px"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "a83c5cb9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "N of observations: 19,859,174\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>userid</th>\n",
       "      <th>day</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>touches</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:00</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:03</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:06</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   userid         day            timestamp  touches\n",
       "0       1  2020-11-22  2020-11-22 00:00:00        6\n",
       "1       1  2020-11-22  2020-11-22 00:00:03        1\n",
       "2       1  2020-11-22  2020-11-22 00:00:06        2"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "by_second_unclean = pd.read_csv(\"Processed Data/touch_sensor_unclean.csv\") \n",
    "by_second_unclean = by_second_unclean.drop('Unnamed: 0', axis='columns')\n",
    "print(f\"N of observations: {by_second_unclean.shape[0]:,}\")\n",
    "by_second_unclean.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2666ac1e",
   "metadata": {},
   "source": [
    "A brief, general look at statistical outliers:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "ce751cb3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "At least one outlier belongs to: 128 out of 201 of the users\n",
      "\u001b[1mOutliers are 355,932 out of 19,859,174 = 1.79%\u001b[0m\n",
      "touches\n",
      "1      4266857\n",
      "2      4127960\n",
      "3      2299172\n",
      "4      1728830\n",
      "5      1200586\n",
      "        ...   \n",
      "960          1\n",
      "966          1\n",
      "983          1\n",
      "985          1\n",
      "998          1\n",
      "Name: count, Length: 565, dtype: int64\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fluve/anaconda3/lib/python3.11/site-packages/seaborn/categorical.py:1794: FutureWarning: use_inf_as_na option is deprecated and will be removed in a future version. Convert inf values to NaN before operating instead.\n",
      "  with pd.option_context('mode.use_inf_as_na', True):\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "z = np.abs(stats.zscore(by_second_unclean['touches']))\n",
    "threshold = 3  #this is the typical chosen score for outliers\n",
    "outliers = by_second_unclean[z > threshold]\n",
    "#print(\"User IDs with outlier observations:\\n\", np.sort(outliers.userid.unique()))\n",
    "print(f\"At least one outlier belongs to: {outliers.userid.unique().shape[0]:,} out of {by_second_unclean.userid.unique().shape[0]:,} of the users\")    #should be 1.79%\n",
    "print(color.BOLD + f\"Outliers are {outliers.shape[0]:,} out of {by_second_unclean.shape[0]:,} = {str(round(outliers.shape[0]/ by_second_unclean.shape[0]*100, 2))}%\" + color.END)\n",
    "print(by_second_unclean.touches.value_counts().sort_index())\n",
    "sns.boxenplot(data = by_second_unclean, x=by_second_unclean.touches, color='#4CC9F0')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "97e0db93",
   "metadata": {},
   "source": [
    "Now, treating those statistical outliers. **Run the following cell only once.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "1964d640",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_29266/1054526867.py:6: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  by_second.touches[outliers] = np.nan\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>userid</th>\n",
       "      <th>day</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>touches</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:00</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:03</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:06</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   userid         day            timestamp  touches\n",
       "0       1  2020-11-22  2020-11-22 00:00:00      6.0\n",
       "1       1  2020-11-22  2020-11-22 00:00:03      1.0\n",
       "2       1  2020-11-22  2020-11-22 00:00:06      2.0"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Replacing the 1.79% outliers for the data that has seconds as units with their median\n",
    "median, std = by_second_unclean['touches'].median(), by_second_unclean['touches'].std()\n",
    "outliers = (by_second_unclean['touches'] - median).abs() > std\n",
    "\n",
    "by_second = by_second_unclean    \n",
    "by_second.touches[outliers] = np.nan\n",
    "by_second['touches'].fillna(median, inplace=True)   # by_second is the cleaner version\n",
    "\n",
    "by_second.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "f1bc6cc6",
   "metadata": {},
   "outputs": [],
   "source": [
    "by_second.to_csv('Processed Data/by_second_clean.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "60460d95",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fluve/anaconda3/lib/python3.11/site-packages/seaborn/categorical.py:1794: FutureWarning: use_inf_as_na option is deprecated and will be removed in a future version. Convert inf values to NaN before operating instead.\n",
      "  with pd.option_context('mode.use_inf_as_na', True):\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxenplot(data = by_second, x=by_second.touches, color='#4CC9F0', \n",
    "              k_depth='trustworthy', trust_alpha=0.05)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "900a53f8",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fluve/anaconda3/lib/python3.11/site-packages/seaborn/_oldcore.py:1119: FutureWarning: use_inf_as_na option is deprecated and will be removed in a future version. Convert inf values to NaN before operating instead.\n",
      "  with pd.option_context('mode.use_inf_as_na', True):\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkwAAAHpCAYAAABwXWMDAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAABW3ElEQVR4nO3dd3QUZf/+8WvTA4RQQggBQugdRHrvvQoq8ACGZgUFQQQeQQRUBBVBQbCC+qhYsdMFBKVDlBKqNKkGDCEEEpLcvz/4Zn/ZbJIJIbArvF/n7DnZqZ+dmb33ysy9szZjjBEAAAAy5eHqAgAAANwdgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMCCWwWmuLg4vfLKK2rWrJmCgoLk7e2twMBAhYeHq169eho0aJBmzpypY8eOubrUf7WFCxfKZrPZH88995yrS8IdKO0xGB4e7upy7mjh4eEO+wP4t7kVn2tuE5gOHDigatWqacyYMVq3bp3OnTunpKQkxcbG6ujRo9qyZYsWLlyo0aNH65dffnF1uTfFc88957DDFy5c6OqSgH+lI0eOOLyXWrRo4eqSbjm2AZC7vFxdgCQZY9SnTx8dPXrUPiwoKEg1a9ZUvnz5dO7cOe3Zs0fnz593YZUAclOvXr3sfwcHB7uwEnTq1Elnz551dRmAW3OLwBQZGant27fbn3fv3l1ffvmlvLy8nKb7/PPPFRQUdKtLBJDLvvzyS1eXgP/z5ptvuroEwO25xSW5/fv3Ozxv3ry5U1iSpLvuuksvvviiOnTo4DA8o1PPMTExGj16tEqVKiU/Pz9VqFBB06ZNU1JSkn2d/fr1U3BwsPz8/FStWjXNnj1bxhin9f7www8aNmyYmjRpovDwcAUGBsrb21sFCxZU7dq1NXr0aP35558Zvrb0fQOMMXrnnXdUv3595c+fXzabTQMHDpTNZtPkyZMd5h00aFCGl+jWrFnjMHzgwIG6cOGCxo4dq3LlysnPz08hISGKiIjQ4cOHs7UPzp8/r9GjR6tMmTLy9fVVSEiIBg0apBMnTmQ6z969ezVixAjVrFlTgYGB8vHxUXBwsFq1aqXZs2crLi7OaZ6M9lVCQoJeeeUV1axZU/7+/goMDFSHDh20cePGTNcdExOjl19+Wc2bN7f3dytUqJCaNGmi1157TZcuXcrW605v2bJluv/++1WmTBnlyZNHPj4+CgkJUY0aNTRgwADNnj1bFy9edJrv6tWr+vjjj9WjRw+FhYXJ399fefPmVXh4uO655x599tlnDtNn57g4cuSIfXpjjH744Qfdf//9Cg8Pl7+/v/LkyaOKFSvq0Ucf1d69e7N8XevWrVNERITKly+vfPnyyc/PT6VLl1ZERIS2bNmS4Typx2XqY82aNdqxY4fuv/9+BQcHy9fXV+XKldPEiROVkJBw3ds6qz5MGfVHOH36tEaMGKHSpUtneoymHl+lS5d2WN7atWuzvDyV0+3bokULh+UeOXJEX331lVq0aKECBQrYt5skrV+/Xk8++aRatmypsmXLqmDBgvLy8lJgYKCqV6+uRx99VL///nuW22z79u169NFHVb16dRUoUMB+fDZq1EjPPPOM4uLicrQNstOH6dy5c3rxxRfVpEkT+3uuYMGCqlOnjsaPH6/jx49nOF9Gy/7qq6/UsmVLBQYGyt/fX7Vr19ZHH32U5WvPSG61hadOndKkSZPUoEEDFSpUSN7e3goKClKbNm303nvv6erVq9la99mzZ/X444+rdOnS8vHxua7LoIsWLVLXrl1VsmRJ+fn5ydfXV6Ghoapdu7aGDh2q+fPnKzk52Wm+G2kLr7fdkqSUlBR9/fXX6tWrl32ePHnyqEyZMvrPf/6jlStXZriujLqcHDx4UIMHD1bx4sXl4+OjsLAwPfHEE7pw4UKGy4iPj9dzzz2nChUq2NuABx54INPP37Ry5RK1cQNff/21kWR/FClSxLzxxhvmwIED2Zr/8OHDDvNXqVLFlC9f3mFY6uO+++4z69atM/ny5ctw/KhRo5yW37lz5wynTfvw9/c3S5YscZq3VKlSDtMNGDDAad6IiAjL5UsyCxYsMMYYs3r1aofhrVq1MmXKlMlwnoIFC5pt27Y51LRgwQKHae69914TGhqa4fylSpUy//zzj9PreuWVV4yXl1eW9YaHh5vIyMgs91W1atXM3XffneH8vr6+ZuPGjU7rXrdunQkJCcly3eXLlzf79u3L1vGT6uWXX87Wfti5c6fDfAcOHDA1atTIcp7mzZtf93Fx+PBhY4wxsbGxpmPHjlku39vb28yfP9/pNV29etUMGjQoy3ltNpuZOHGi07zpj8t+/foZT0/PDJfRo0eP69rWxhin4yyt9Mdop06dTOHChS2P0fTHV3b2x41s3+bNm1vux9WrVxtjjBk2bJhlXZ6enua9995zWk9ycrIZPny45fyHDx/O0TZIfzymt3LlShMUFJTl8vLkyWM+/vhjp3nTL/uBBx7IdBmvvfaa5XGT1o22hcZc+/zJnz9/lq+tXr165vTp01muu2XLlqZEiRJZvu8zk51jQ5K5ePGiw3w30hbmpN06f/68admypWWdvXv3NgkJCQ7zTpo0yWGae++91/j7+2c4f926dU1iYqLD/BcuXDB16tTJcPqAgADz8MMPOwybNGmSw/zp3xfZ3TdpuUVgOnnyZKYfvgUKFDCtWrUykyZNMn/88UeG82fWQNSoUcO0aNHC2Gw2pze2zWYz9evXd9oBnp6e5vjx4w7L79y5s/H29jbVq1c3LVq0MN27dzft27c3pUuXdpg3JCTEXL582WHe9I2FdC0I1KtXz3To0MEULVrUTJ8+3fTq1ctUrlzZYbo6deqYXr162R+pDW/6N2rqo2bNmqZVq1ZOYbBMmTLmypUr9prSfxilPmrVqmWaNm3q9KH4wgsvOLymjz76yGneypUrm7Zt2zp9qIWEhJjo6GjLfRUeHm7atm3r1HC1bdvWYd0HDx50mqZatWqmS5cupmrVqk6v+9KlS9k6BhMTEx22m4+Pj2natKnp1q2badCggUNDmDYwXbhwwWkf22w2e00NGjQw3t7eloEpo+MiNTClD+xFihQxHTp0MC1btjQ+Pj4O6/3pp58c1vPYY485NSxt2rQx7dq1czpO5s2b5zBvRkHe19fXNGvWzFSvXt1p3K+//pqtbZ0q7bxWgSm7x+jZs2dNr169nAJQUFCQw3vp2Wefta/rRrZv+sCU2obUqlXLdOrUyZQqVcohMHl4eJhKlSrZj62OHTs6ve/9/PzMyZMnHdYzcuRIp/WEhITY92VqmDl8+HCOtkFWgSkqKsrkzZvXYXxoaKhp3769Uzjx9PQ0a9ascZg/o2O9UKFCpm3btk4BIzAwMNvvWWNuvC389ddfjbe3t8M+rlOnjunSpYspW7asw7wNGzY0KSkplusODg427dq1M02aNHFqvzJy4sQJh8+ovHnzmlatWpmuXbuaOnXqmODgYPu4tIHpRtrCnLZbrVu3djpWmzVrZho0aOD0GT506FCHedMHptTjpX79+qZ+/fpO49KH76FDhzrVW7duXdOsWTPj5+fnNP9tG5iMMebZZ5/N8OBL/+jatas5e/asw7wZfQin/Y95zJgxTuPff/99+/ju3bs7jPvggw8clr9nz55M38RPPfWUw7zpzzKlPyhLlSpl9uzZYx+flJRkkpKSjDHOB1TqGaX0Mnqjzpkzxz7+zz//NMWKFcv0NWX0YZR2XenHt2zZ0j4uOTnZ6WzUiy++aB9//vx5pxA6bty4LPfV4MGD7dtg7969Dh9SPj4+Dv9p9O/f32HeTz/91GHbvPjiiw7jX3nllQy3YXonTpxwmO/DDz90mubIkSPm7bffNqdOnbIPS3/cBgcHm99++81hvujoaLNo0SKHYdk9LlauXOkwXbdu3Rz+c9u3b5/Dh0K1atUcxnl4eNjH1atXz1y4cME+/syZM6ZkyZL28YULF3ZYdvrAFBgY6HDGMP34yZMnZ2tbp0r/+tO6kWPUmOw3jjeyfY1xDkwFChQw69evt49PSUmxL+/AgQMmJiYmwzrmzJnjsJy04fXAgQNOAXHy5Mnm6tWr9mmSkpLMl19+meU/J1l9QGQVmPr06eO0jVL/MUxOTjYPPfSQw/gGDRpkuey7777bnDt3zhhjzMWLF50+3NeuXZtpnendaFvYpEkT+3AvLy/zyy+/2MelpKQ4nbX48ssvs1z3gAEDHAJZ2r8z8+uvvzosI20NqaKioszs2bMdjs0baQtz0m4tXbrUYZ6CBQua3bt3O2yPtMepzWYzUVFR9vHpP988PT3NypUrMx0/aNAg+7hTp045BbK0+2LHjh1OZ6tu68BkjDHvv/9+hv+NpH+kT/rpN0S+fPlMfHy8ffw333zjML5s2bIO63399dcdxqcNAMYYk5CQYN555x3TqVMnExYWlulpRMn5lHL61/PRRx9l+vpzGpjKlSvnsD2MMeb55593mKZ///72cek/bOrXr+8w77lz5xzGV6hQwT5u8+bNDuOKFy9uDzupfvzxR4dpqlatah+Xfl/5+fmZ8+fPO8yf/uxF6n/bycnJDv9R+fj4OPzH3KtXL9OuXbscvSkSEhIc/ouuV6+emTdvnlmxYoU5cuSI0/ZNVa1aNYf1LVy4MFvry+5xkf5UfaNGjZxec8GCBR2mST0zlf4S41133eU0b/r/8FPPhhjjHIjGjh3rUNtXX33lMP6hhx7K1mtPlXZeq8B0PceoMdlvHG9k+xrjHJimTp2a6etNTk42n332menZs6cpU6aM/Ux3Ru3IiBEj7PO98sorDuNatGiRre2bG4EpOTnZBAQEOIxL+yFojDH//POP05m4tP/Upl92+n8qH3/8cYfxn3zySbZenzE31haePXvWYfsXKFDAad83btzYYd6IiIhM112wYEETGxub7dpT/fXXXw7LadeunXn33XfNmjVrzIkTJzKc50bbwpy0W+nPVqdvD4wx5r777nOY5uWXX7aPS//51rt3b4d5t23b5rQdUn366acO49KHcmOcz0ClD0y5wS2+JZdq0KBBGjhwoDZt2qS1a9dqw4YNWrdundPtBDZs2KANGzaoUaNGGS6nbNmy8vf3tz8PCAhwGF+1alWH5+nHp+3AevnyZbVs2VKbNm3K1mvIrLNaqptxL5Tq1as7ddSsVq2aw/O0t2xIr27dug7PAwMDHZ6n3R5pOyJLUuXKleXp6ekwrGbNmg7Ps+psWa5cORUsWDBb6z937pxiY2PtwxMTE/XVV19lumyrdafl4+OjiRMnaty4cZKkzZs3a/Pmzfbx+fPnV7NmzfTQQw+pa9eu9uHpOxs2b948W+tLL7PjIn39v/32m+WyDh8+rPDwcKd5IyMjFRkZaTlvZrVcz3GS227Wum9k+2Yks21njFGvXr30zTffZKuutO1Ibh1jOXHu3DmHLzn4+PioYsWKDtMUKFBAYWFhOnjwoKRrr/XIkSMqUqRIhsu8mcfR9bSFR44ccfiST0xMzA21J3fffbfTZ0l2FC9eXI888ojmz58vSVq+fLmWL19uHx8UFKRWrVpp+PDhatq0qaQbbwtzckylb/urV6/uNE3NmjX1xRdfZLjO9K7nOEj/+ZXRutPv55vBLb4ll5bNZlODBg00duxYffPNN/r777/13XffKV++fA7TRUVFZbqMAgUKODz38HB8mek/oLMyd+5ch7Bks9lUp04d9ejRQ7169VKdOnUcpk/7BsxIaGhottd9qxQuXNjhefoAlFb613ejdwVOv26r9V+v6/m23NixY7Vq1Sr169dPpUqVcnhtsbGx+uGHH9StWze9/vrruVZfqtw8LnL6DUGrea/nOMltrlx3ellto8z241dffeUUlqpXr65u3bqpV69eatasmcM4q3bkVrkZdbjTvrxeOdn32TFv3jx99dVX6tmzp4oVK+YwLjo6Wp9//rmaN2+e7cCdkRtpF6Sb3/b/G44DtwhMFy5cUHx8fIbjPDw81LVrV7Vt29ZhuLe3960oTevWrXN4vmjRIm3ZskWLFy/Wl19+qZ49e17X8tKHt7RyegDu2rXLadju3bsdnpcqVSpHy04v/VeV9+zZ4/RV1z/++CPLeXKqcOHCDv/B5c+fXwkJCTLXLi1n+IiOjr6udbRq1Ur/+9//dOTIEV26dEn79u3TggULHAL7zJkz7X+XKVPGYf61a9fm6LVldlyk33aLFi3K8vUaY9SlS5cM533ppZcs5x0+fHiO6nc32X0v3cj2zUhm+zF9OzJ9+nT98ccf+vbbb/Xll1/qkUceyXSZOT3GcuMnToKCghyO/cTERKfbwMTExDj8XJUrf+rmetrC9P8UVapUyXLfb926NdN1Z9W2Z0fPnj311Vdf6eTJk4qLi9OuXbs0e/Zse5AwxmjWrFmSbrwtzMkxlf69snPnTqdpblbbHxYW5vA8O/v5ZnCLwLRz506FhYXpv//9b4Yb4tixY0735El/We1mSX//jTx58tj/3r9/v2bPnp1r60p7GVFSlvdASuvAgQP207nStdOXc+fOdZimTZs2N16grp12Tvsf0IkTJ/Tqq6/an8fExDj9hk9WHzDXw8PDw2FZsbGxGjVqlNMpfGOMNm3apJEjR2rx4sXZXv6LL76ozZs32/+T8vf3V4UKFdS3b1+HO1GfPn3a/nePHj0clvH0009rw4YNDsNiYmIcTlNfj27dujk8nzhxYoanuU+cOKG5c+fq8ccftw/r0qWLwwfCq6++6nCD2FTR0dFauHCh/vOf/+SoRneU/r108uTJDKe7ke17PbJqR06fPq3nn38+03m7devm8GG8Zs0aTZkyxX5POenaMf/tt986fChmdxtkxcPDQ506dXIYNm7cOPt7LiUlRePHj1diYqJ9fL169TK9HHezXU9bGBwcrAYNGtiH7927Vy+99JLTP4BJSUlavXq1hgwZku2uGdcjPj5eL7zwgsNnX968eVW1alUNGDBAfn5+9uGpbc+NtoU5abfSt+Nvv/22w/3J1q1bp6+//tr+3GazqXPnzpavPztatGjhcG/GDRs2OJxt++OPP/Txxx9nuYzb5j5M69atc+isFRQUZJo3b266detmmjRp4vC1T+naV4uz6vSdvnNj+s55aTvuGePcwTRtZ7HJkyc7jPPx8TGtW7c2zZs3Nz4+Pk6dNtN3NLO6v0la3377rcO0vr6+pm3btvZOfKnfTMns66x33XWXad26tVMnzdKlS2d5W4GMOselHZ+dbzBVqVLF4evNqY/g4GCHDqDZ6YiaviNt2k626b+1JF37inLLli1Nt27dTKNGjUxgYKB9XGYd5zOSOl/hwoVN48aNTbdu3UynTp2cvmVz11132ef5559/HL5ppv/r9Fq9enXTpUsX06hRI+Pn52d5W4GstG3b1mFaT09PU7duXdOtWzfTpk0bEx4enun2fPDBB532Vc2aNU3Xrl1Nu3btTIUKFezfpEu/n9N3+k7bIdwY6/eVles5xq73GDXGmEKFCjm97p49e5pevXo5dDy+ke2b1bGa1gcffOAwnYeHh2natKlp06aNyZcvn1M7kn5bpu8YLckUK1bMtG3b1rRv395+L57068/uNsjqeNy9e7fJkyePw/jMbivg4eFhfv75Z4f5rY717H7ZJSM32hauXbvW6dtXqdu1S5cupm7dug6vPe174EaP/1T//POPfRkhISGmefPm9lvXpN9/ae93diNtYU7brRYtWjjM4+/vb5o1a2YaNmzotB3TfsvNGOv9bPXZMHjwYKdjrV69eqZ58+Z31m0F1q9fn+FBn9EjLCzM6SZcNzMwnT9/3ul+HKmPwoULm3HjxmW5k67ng/Hy5csmLCws09eeeg+O9K+nU6dOTl/NTX0EBgaaLVu2ZPv1pko7PqMPo5deeinTmxim3VfpbxR3o4HJGGPWrFljebO21EdW30pML23jktnD39/frFq1ymG+ffv2Zbr9M3ud13NcXLhwwbRv3z5br7d169YO8yYmJmZ5o8C0j/TfHv23B6aMbieS+njjjTfs093I9s1uYEpMTMzwXjOpx9TUqVOz3JZJSUnmkUcesawv/fqzuw2sjsdly5Y5fXhn9Doyuh3HrQxM19sWGmPM559/bnnjytTHunXrMl13bgSmrB6FCxc2u3btcpj3RtrCnLRb586dM82aNbNcV69evZxuqXCjgSkmJibTmxz7+fmZvn37Ztlm5EZgcotLco0bN1ZkZKRmzJihXr16qWrVqipQoIC8vLzk4+OjokWLqlWrVpo5c6Z2796tChUq3LLaChYsqA0bNujhhx9WaGiovL29FRoaqoEDByoyMtLpGyM3ws/PTz///LP69OmjkJCQbHeCK1KkiDZu3Kj//ve/KleunP0nSvr376/t27c7dUzPDWPHjtUff/yh4cOHq1q1agoICJCXl5eCgoLUvHlzzZw5U7t27dLdd9+d6+tu3ry59u7dq9dee02tW7dWcHCwvL295evrq+LFi6tly5Z65plntHHjRvXv3z/by/3oo480ZswYNW3aVOHh4QoICJCnp6cCAwN11113aeTIkdq5c6datWrlMF+FChW0fft2ffDBB+ratatKlCghX19f5cmTR6VKlVL37t2z7KNiJX/+/Fq6dKl+/PFH/ec//1HZsmWVJ08eeXp6qmDBgqpVq5aGDBmiRYsW6bvvvnOY19vbWx988IHWr1+vwYMHq3LlysqXL588PT2VP39+VatWTf3799f777+f6U+k/Fu98MILev7551WlShWHyxrp3cj2zS5vb2+tWrVKTz/9tMLDw+Xt7a0iRYro3nvv1ZYtW9SkSZMs5/f09NS8efO0efNmPfzww6pataoCAgLk7e2tokWLqkGDBho/frzT72xmdxtYadeunfbu3aupU6eqYcOG9p91yZ8/v+6++249/fTTioqK0oABA3K8jtyQk7bwvvvu0759+zRlyhQ1adJEhQsXlpeXl/z8/FSqVCm1b99eU6dO1c6dOy33U04EBATo008/1eOPP64GDRooLCxMefPmlZeXlwoVKqR69erpmWee0a5du5y6otxIW5iTdqtQoUJavXq1Pv/8c/Xo0cM+j5+fn8LDw9W7d28tXbpUX375pXx9fXN1OwUGBuqXX37RxIkTHfbt/fffr23btqldu3a5ur6M2Ixxk69jINvWrFmjli1b2p9HRETYf2cOAO4UtIW4ldziDBMAAIA7IzABAABYIDABAABYoA8TAACABc4wAQAAWCAwAQAAWCAwAQAAWCAwAQAAWCAwAQAAWCAwAQAAWCAwAQAAWCAw3WJ9+vTRgAEDNGTIEPXr10/PPPOMdu/ebR//3XffadGiRVkuY/369dqzZ0+W0yxYsEArV66UJC1cuFBz5sy57loPHjyon3/+2WHY0KFDFR8ff93LyomTJ0/qoYce0oMPPqglS5Y4jDt9+nSOfwg1u/r06aODBw/e1HXcCbp3767Tp0/neP7Tp0+rS5cuGY6Li4vTJ598ku1ljRs3TsePH89xLRmZNGmSli5d6jQ8q7r/TdK2Jbej7LSnNyIyMlJDhw69acvPqet974DA5BKTJk3Se++9p48//lgdOnTQ2LFjFRUVJUnq1q2b+vTpk+X8Vm/w5ORkDRo0SG3atLmhOjMKTO+++67y5MlzQ8vNrrVr16pixYp655131LFjR4dxtyIwubvk5GRXl+By19vov/TSSypZsuRNrMh1btbxkBttya2WkpKS7WlvdmByVwSm6+fl6gLudE2bNlVUVJQ+++wzPffcc1q4cKHi4uI0fPhw7dmzR7NmzVJKSoqSk5PVo0cPhYSE6LffftPWrVu1dOlS3XPPPSpevLhmzZqlKlWqaP/+/erfv782btyocuXK6d5775UknT17VqNGjVJ0dLRKlCihcePGKX/+/A7rk6TFixdr3759euSRR7RgwQLFxcVp6NChqlKlikaNGqWWLVvq+++/V758+bRv3z698cYbunz5snx8fDRs2DBVq1ZNp0+f1tChQ9WrVy9t2LBBly5d0hNPPKH69es7vf7Lly/rjTfesAfGFi1aKCIiQsuXL9cXX3yhlJQURUVFaeLEiSpVqpR9vpkzZ+rMmTMaOnSoihYtqhdeeMGynh9++MG+zk6dOmn16tWSpN27d2v+/Pn2M2eDBw9W48aNJUnr1q3Ta6+9pvPnz6tTp04aMGCAJOn8+fN64403dPr0aSUkJKhx48YaMmSIjDF6/fXXtX37dnl5ecnT01Nz5syRj4+Pw+t+6aWX5OHhoWPHjunChQv27evr66v4+Hi9+eabOnTokBITE1WlShWNGDFCXl5eGjlypMqWLau9e/fK19dXM2fOdFjujz/+qC+++EJeXl5KSUnRmDFjVLlyZf3111+aM2eOYmJidPXqVXXp0kX33HNPlq8/p/v3t99+01tvvSUvLy/Vq1cv2++FzNaXat68edq6dauSk5P1+OOPq3bt2po5c6bi4+M1dOhQeXp66q233tIXX3yhVatWKSkpSV5eXnr88cdVtWpVSdfOGj7//PMqV66cRo4cqYoVKyoqKkrR0dGqU6eORo0aJUlZ7oNjx45pxowZiouLU4kSJZSQkJDl68qo7tmzZysoKEj9+vWTJB0/flyjR4/Wp59+Kk9PT4f5f/31V7399tv27fnTTz/prbfeUkhIiPr06aOWLVsqMjJSxYsX1+jRozN8P0nSyJEjde+996pJkyaSrv3j1rBhQ3Xo0CHL4/Gll16ytyULFy7U0aNHlZCQoJMnT6pQoUKaPHmyAgIClJSUpDlz5mjbtm0KCAhQtWrVtH//fs2aNctpm2T2/lm5cqVWrVqladOmSZKMMerXr5+mTp2qsmXLavny5Vq8eLGSk5Pl7++vJ554QmXLltXSpUu1bNkyBQQE6K+//tLIkSP1yiuv6IMPPpDNZpMkDR8+XAMGDHBohzZt2uTUnnbu3FmLFi3S0qVL5eHhoTJlyujJJ59U3rx5M20vx40bJ0n69NNPtWLFCtlsNvn6+uq1116TdC3Mzpo1Szt37lRycrLGjx+vihUrOm2XzI67vXv3aubMmXr//fft044cOVL33XefGjdurC1btujDDz9UQkKCPD099dBDD6lWrVqKjIzU7NmzVbNmTad1Z/TegQWDW6p3797mwIEDDsN++eUXExERYYwxZsGCBeaNN94wxhjz3//+16xcudI+XWxsrDHGmGnTppkvvvjCPnzHjh2mZcuWJjIy0j4s7TQLFiwwPXr0MOfOnTPGGPPaa6+ZV155xWl9xhjz9ddfm2nTphljjFmyZIl55plnHGpt0aKFuXjxorl69aq5//77zebNm40xxvzxxx/mnnvuMfHx8ebUqVOmRYsWZu3atcYYYzZt2mT69++f4faYP3++mTp1qklJSTGXL182Q4cONT///HOGtaW1Y8cOM2TIEPtzq3o6d+5snzY+Pt60aNHCvk179Ohhfv/9d2OMMSkpKfbt3Lt3b/P6668bY4yJiYkxnTp1Mn///bcxxpinnnrKvr2TkpLMmDFjzOrVq83+/fvNAw88YFJSUowxxsTFxdn/TmvatGkmIiLCXLp0ySQnJ5vx48eb//3vf8YYY1555RWzbNkyez0zZswwn376qTHGmBEjRpgxY8aYq1evZrhdOnXqZKKjo+3bJD4+3iQnJ5uHHnrIHD161BhjzJUrV8zgwYNNVFRUpq8/p/v3n3/+MV27djVHjhwxxhjz/fffmxYtWphTp05lWG9291+LFi3Mjz/+aIwxZvfu3aZHjx7m0qVLTvs2tYZUu3fvNgMGDLA/T/v+GzFihJkwYYJJSkoyV65cMX369DG7du2y3AcPP/ywvZZDhw6ZNm3amCVLlji9pqzqPnbsmOndu7dJTk42xhjz+uuvmw8++MBpGanbM3Xf/fTTTw7bs3fv3mbGjBn2Yyyr99OIESPMunXr7Mt+9tln7XVndTymb0t69+5tLly4YIwxZvLkyfbpFi9ebEaPHm2uXr1qrl69akaPHm1GjBjh9JqMyfz9c+XKFdOtWzd7W7V9+3bz4IMPGmOM2blzpxk7dqxJTEw0xhjz+++/29vNJUuWmPbt25tjx47Z1zFs2DCzZcsWY4wx+/fvN/369cv0vZi2Pd24caN54IEHzMWLF40x146FmTNn2l9/Zu3l0qVLzSOPPGLi4uKMMdfaluTkZLNjxw7TqlUrs2fPHmOMMd9++6156qmnMtwuWR13/fv3N3v37jXGGHPy5EnTs2dPk5SUZE6ePGkee+wxc+nSJWOMMX/99Zfp2bOnSUxMzHLdGb13kDXOMLmxWrVq6aOPPtKJEydUq1YtVa9ePdNpixUrppo1a2Y6vkGDBipUqJAkqUuXLnr22WdvqLZjx47JZrOpbt26kqTq1aurYMGCOnjwoIoUKSIfHx81bdpUklS1alWdPHkyw+Vs27ZNjz32mGw2m/z8/NSuXTtt3bpVLVu2zNV6MrN7926FhYWpRo0akiSbzaaAgAD7+NatW0uSAgMDFRoaqlOnTilfvnzavn27/vnnH/t0ly9f1vHjx1W3bl0lJydr+vTpqlWrlho0aGD/Dze9li1b2i9vdurUSV9//bX69eun9evXa/fu3fr8888lSQkJCfLw+P9Xz9u2bSsvr4zfurVr19aLL76oRo0aqX79+ipRooSOHj2qI0eOaMqUKfbp4uPjdfToUcXExGT4+v/8888c7d89e/aobNmy9rOBnTp10uzZszPd/qms9p+np6c6dOggSapSpYoKFy6sgwcPKjg42GlZBw8e1EcffaTY2Fh5enrq+PHjSkhIkK+vb4b7wNPTU56enipXrpxOnjypqlWrZroP4uPjdfDgQXstZcqUyfJ9mVndNWrUUHh4uH799VfVrVtXP//8s8PZg1Sp2zMsLEyS1KFDB6ezih06dLAfYzfyfsrseEyvXr16yp8/v6Rr+/7PP/+UJG3fvt3h2Gzfvr1+/PFHp/mvXLmS6funRYsWatasmZYvX64+ffpo6dKl9svxv/76qw4ePKhHH33UPt/FixftZ/iqVq3qcLn13nvv1eLFi1WnTh1988036t69e6bvxbS2bdumli1bKl++fJKudZN47rnnLOfbsGGDunXrprx580qSQztSvHhxVa5c2V7nZ599luEysnrvd+jQQUuWLFHFihW1dOlStWnTRp6entq8ebNOnDihJ554wr4cm82mM2fOXNe6YY3A5Ab27t2r0qVLOw2/99571bhxY23btk3vvvuuSpcurZEjR2a4DH9//xyt29PT0+F6f2JiYo6WI8mhMfL29rY/9/DwyHafguw0aNdbz428xrSX0jw8PJScnCzzf79X/eabbzpdapOudZL9/ffftWPHDr3zzjuaPXu2ihcvnu11GmM0ZcoUlShRIsPxWe3ryZMna//+/YqMjNTYsWM1ZMgQlSlTRgEBAXr33Xedpt+4cWO268rJ/r2R/ZmTeZOSkjRx4kS99tprqlSpkuLj49W5c2ddvXo1w8CU0f6VMt8HGX3hIaevsVevXvr0008VExOj2rVrq2DBgjlaTlbHQ9racuu9ntk2y2rdaVm9fzp16qTp06ere/fu2rBhg/3ylzFGHTp0yLQDdfrt0LRpU82fP18HDhzQb7/95hC0rkdubMPsbrOs3vsdOnTQkCFD9Oijj2r58uUOly3r1KmjCRMmOM0THR2d7XXDGp2+XezXX3/Vt99+q/vvv99p3PHjx1WsWDF16dJF/fr1s3dMzJs3ry5dunRd69m4caP9P7off/xRtWvXlnTtv499+/YpJSVFCQkJ+uWXX+zzZLWesLAwGWO0detWSdfO1Jw/f17lypW7rrpq166tn376ScYYXblyRcuXL7efZchK+tqyqqdQoUIyxujo0aOSpOXLl9vnq1q1qv766y/98ccfkq41PhcvXsxy3f7+/qpVq5ZDh8lz587p77//VkxMjK5cuaI6depo6NChCgkJsa83vbVr1+ry5ctKSUnRkiVL7PukSZMm+uSTT+wN28WLF3XixAnLbZKcnKyTJ0+qYsWK6t27t5o3b669e/eqZMmSyps3r8M3DU+cOKGLFy9m+vpzun+rVKmiQ4cO6dixY5Kkn376SUlJSfbx77zzjhYvXuw0n9X6kpOT7ftt7969OnfunMqVK6e8efMqISHBvo7ExEQlJSWpaNGikqSvv/7acrtlJLN9kCdPHpUvX17Lli2TJB05ckQ7d+7MdDmZ1S1JderU0fnz5/XRRx/Z+5Oll7o9U7/Zt2LFCoftmV5W76fixYvb+zadOnXKqe7MjsfsqlWrllauXKmkpCQlJSXZt1F6Wb1/JNnPhsybN0+1a9e2n6lp1KiRVqxYobNnz0q6dqzu27cv03o8PT3VrVs3PfPMM2rSpIn9jFF66duS2rVra82aNfZw/P333ztsw8zay0aNGum7776zLysuLu66Op9LWb/3CxcurEqVKmnu3LkqUKCAwsPDJUl169bVtm3b7Gf6pGvHmpX07x1Y4wyTC0yePFk+Pj66cuWKwsPDNX36dHsjkdbixYu1fft2eXt7y8PDw/4fUtu2bfXSSy9p/fr16tGjR7bOXtSoUUNTp0516PQtSc2aNdOaNWsUERGhIkWKqHz58rpy5Yok6e6779Znn32mIUOGqGrVqvYOsZLk5eWlKVOm6I033tC8efPk4+OjyZMny9/fXxcuXMj2tnjggQf0xhtvaPDgwZKudVJt0aKF5XxlypRReHi4Bg0apNDQUL3wwguZ1iNJTzzxhMaPH6/8+fOrefPm9uUEBARo6tSpevPNN3X58mXZbDYNHjxYjRo1ynL9EyZM0Ny5czVo0CD75Y/Ro0crJSVFL7/8spKTk5WcnKzq1atn2vG5UqVKGjNmjL2TbWoH/WHDhuntt9/W0KFD5eHhIU9PTz388MOW+zklJUUzZsywX4oqUKCAxo4dK09PT02bNk1z5szRl19+qeTkZAUGBmrixIkKCgrK9PXnZP+mrnPixIn2Tsqpl28k6dChQ6pQoYLTfFbHU968eXX48GENGTJEycnJmjBhgv3yUfv27TV48GD5+/vrrbfe0pAhQ/TII48oMDBQrVq1ynKbZSarfTB+/HhNnz5dn3/+uUqUKJHlpfCs6rbZbOrUqZNWrlxp75Se0fYcM2aMJkyYIG9vb9WpU0f+/v6Zfvhn9X7q27evJk+erMGDBys8PNypzcnseMyurl276s8//9TAgQMVEBCgihUrKjo6OsNpM3v/pF4+79Chg9566y1Nnz7dPk+NGjX08MMPa8KECUpOTlZSUpIaNGiQYefpVJ06ddK7776baSCVnNvTzp076/Dhw3rsscccOn1LWbeX7dq107lz5zRs2DB5enrK399fr7766nVtQ6v3focOHTR58mR7PdK1EDdhwgS9+uqrunLlipKSklS+fPkMzzilFRAQ4PTe+e6773Tu3DkNGjTouuq+U9hM6vlRALdM2m8e3SlSUlL02GOPad68ebl66fXfbPz48WrZsqXatWuX6TTx8fH2kLV+/Xq98847+uCDD3K1jtw6HlNrTUpK0gsvvKAKFSqob9++uVTl9Vu7dq2+/fZbp35fQE5whgnALeHh4aH58+e7ugy3sG/fPk2ZMkXh4eGW9zhavHixfv75Z6WkpChv3ryWZw5cafTo0bp69aoSExNVvXp19ezZ02W1PP300/rrr780depUl9WA2wtnmAAAACzQ6RsAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMCCl6sLuNlSUlJ08uRJBQQEyGazubocAACQDcYYXbx4UaGhofLwcP35nds+MJ08eVIlS5Z0dRkAACAHjh8/rhIlSri6jNs/MAUEBEi6tsHz58/v4moAAEB2xMbGqmTJkvbPcVe77QNT6mW4/PnzE5gAAPiXcZfuNK6/KAgAAODmCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWvFxdAHDs2DFFR0e7tIagoCCFhYW5tAYAgPsiMMGljh07pkqVK+tyfLxL6/DPk0d7o6IITQCADBGY4FLR0dG6HB+vrtPfVVDZiq6p4dA+fT92qKKjowlMAIAMEZjgFoLKVlRIlbtcXQYAABmi0zcAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFAhMAAIAFL1cXcCc6duyYoqOjXV2GgoKCFBYW5uoyAABwewSmW+zYsWOqVLmyLsfHu7oU+efJo71RUYQmAAAsEJhusejoaF2Oj1fX6e8qqGxF19VxaJ++HztU0dHRBCYAACwQmFwkqGxFhVS5y9VlAACAbKDTNwAAgAUCEwAAgAUCEwAAgAUCEwAAgAUCEwAAgAUCEwAAgAUCEwAAgAXuw3SHi4qKuqPXDwBAdhCY7lBxf5+RbDb179/f1aVIkhISE11dAgAAmSIw3aGuXIyRjFHbKXNUokpNl9Vx8JflWvf6VCUlJbmsBgAArBCY7nAFw8u79Cdaov/c57J1AwCQXXT6BgAAsEBgAgAAsEBgAgAAsEBgAgAAsEBgAgAAsODSwDRt2jTVrVtXAQEBCg4OVo8ePbRvn+O3pq5cuaJhw4apcOHCypcvn3r16qUzZ864qGIAAHAncmlgWrt2rYYNG6aNGzdqxYoVunr1qtq1a6dLly7Zp3nyySf1/fff64svvtDatWt18uRJ9ezZ04VVAwCAO41L78O0dOlSh+cLFy5UcHCwtm3bpmbNmunChQt677339Mknn6hVq1aSpAULFqhy5crauHGjGjRo4IqyAQDAHcat+jBduHBBklSoUCFJ0rZt23T16lW1adPGPk2lSpUUFhamDRs2ZLiMhIQExcbGOjwAAABuhNsEppSUFI0cOVKNGzdWtWrVJEmnT5+Wj4+PChQo4DBt0aJFdfr06QyXM23aNAUGBtofJUuWvNmlAwCA25zbBKZhw4Zp165dWrRo0Q0tZ/z48bpw4YL9cfz48VyqEAAA3Knc4rfkhg8frh9++EG//PKLSpQoYR8eEhKixMRExcTEOJxlOnPmjEJCQjJclq+vr3x9fW92yQAA4A7i0jNMxhgNHz5cixcv1s8//6zSpUs7jK9du7a8vb21atUq+7B9+/bp2LFjatiw4a0uFwAA3KFceoZp2LBh+uSTT/Ttt98qICDA3i8pMDBQ/v7+CgwM1JAhQzRq1CgVKlRI+fPn1+OPP66GDRvyDTkAAHDLuDQwzZs3T5LUokULh+ELFizQwIEDJUmvvfaaPDw81KtXLyUkJKh9+/Z68803b3GlAADgTubSwGSMsZzGz89Pc+fO1dy5c29BRQAAAM7c5ltyAAAA7orABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYIHABAAAYMGlgemXX35R165dFRoaKpvNpm+++cZh/MCBA2Wz2RweHTp0cE2xAADgjuXSwHTp0iXVrFlTc+fOzXSaDh066NSpU/bHp59+egsrBAAAkLxcufKOHTuqY8eOWU7j6+urkJCQW1QRAACAM7fvw7RmzRoFBwerYsWKevTRR3Xu3Lksp09ISFBsbKzDAwAA4Ea4dWDq0KGDPvzwQ61atUrTp0/X2rVr1bFjRyUnJ2c6z7Rp0xQYGGh/lCxZ8hZWDAAAbkcuvSRnpU+fPva/q1evrho1aqhs2bJas2aNWrduneE848eP16hRo+zPY2NjCU0AAOCGuPUZpvTKlCmjoKAgHTx4MNNpfH19lT9/focHAADAjfhXBaa//vpL586dU7FixVxdCgAAuIO49JJcXFycw9miw4cPKzIyUoUKFVKhQoU0efJk9erVSyEhITp06JCefvpplStXTu3bt3dh1QAA4E7j0sC0detWtWzZ0v48te9RRESE5s2bpz/++EMffPCBYmJiFBoaqnbt2mnq1Kny9fV1VckAAOAO5NLA1KJFCxljMh2/bNmyW1gNAABAxv5VfZgAAABcgcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABggcAEAABgIUeBqUyZMjp37pzT8JiYGJUpU+aGiwIAAHAnOQpMR44cUXJystPwhIQEnThx4oaLAgAAcCde1zPxd999Z/972bJlCgwMtD9PTk7WqlWrFB4enmvFAQAAuIPrCkw9evSQJNlsNkVERDiM8/b2Vnh4uF599dVcKw4AAMAdXFdgSklJkSSVLl1aW7ZsUVBQ0E0pCgAAwJ1cV2BKdfjw4dyuAwAAwG3lKDBJ0qpVq7Rq1SqdPXvWfuYp1fvvv3/DhQEAALiLHAWmyZMna8qUKapTp46KFSsmm82W23UBAAC4jRwFpvnz52vhwoUaMGBAbtcDAADgdnJ0H6bExEQ1atQot2sBAABwSzkKTEOHDtUnn3yS27UAAAC4pRxdkrty5YrefvttrVy5UjVq1JC3t7fD+JkzZ+ZKcQAAAO4gR4Hpjz/+0F133SVJ2rVrl8M4OoADAIDbTY4C0+rVq3O7DgAAALeVoz5MqQ4ePKhly5bp8uXLkiRjTK4UBQAA4E5yFJjOnTun1q1bq0KFCurUqZNOnTolSRoyZIhGjx6dqwUCAAC4Wo4C05NPPilvb28dO3ZMefLksQ/v3bu3li5dmmvFAQAAuIMc9WFavny5li1bphIlSjgML1++vI4ePZorhQEAALiLHJ1hunTpksOZpVTnz5+Xr6/vDRcFAADgTnIUmJo2baoPP/zQ/txmsyklJUUzZsxQy5Ytc604AAAAd5CjS3IzZsxQ69attXXrViUmJurpp5/W7t27df78ef3666+5XSMAAIBL5egMU7Vq1bR//341adJE3bt316VLl9SzZ0/t2LFDZcuWze0aAQAAXCpHZ5gkKTAwUM8880xu1gIAAOCWcnSGacGCBfriiy+chn/xxRf64IMPbrgoAAAAd5KjwDRt2jQFBQU5DQ8ODtaLL754w0UBAAC4kxwFpmPHjql06dJOw0uVKqVjx47dcFEAAADuJEeBKTg4WH/88YfT8N9//12FCxe+4aIAAADcSY4CU9++ffXEE09o9erVSk5OVnJysn7++WeNGDFCffr0ye0aAQAAXCpH35KbOnWqjhw5otatW8vL69oiUlJS9MADD9CHCQAA3HauOzAZY3T69GktXLhQzz//vCIjI+Xv76/q1aurVKlSN6NGAAAAl8pRYCpXrpx2796t8uXLq3z58jejLgAAALdx3X2YPDw8VL58eZ07d+5m1AMAAOB2ctTp+6WXXtKYMWO0a9eu3K4HAADA7eSo0/cDDzyg+Ph41axZUz4+PvL393cYf/78+VwpDgAAwB3kKDDNmjUrl8sAAABwXzkKTBEREbldBwAAgNvKUR8mSTp06JAmTJigvn376uzZs5KkJUuWaPfu3blWHAAAgDvIUWBau3atqlevrk2bNunrr79WXFycpGs/jTJp0qRcLRAAAMDVchSYxo0bp+eff14rVqyQj4+PfXirVq20cePGXCsOAADAHeQoMO3cuVP33HOP0/Dg4GBFR0ffcFEAAADuJEeBqUCBAjp16pTT8B07dqh48eI3XBQAAIA7yVFg6tOnj8aOHavTp0/LZrMpJSVFv/76q5566ik98MADuV0jAACAS+UoML344ouqXLmywsLCFBcXpypVqqhZs2Zq1KiRJkyYkNs1AgAAuNR13YcpJSVFL7/8sr777jslJiZqwIAB6tWrl+Li4lSrVi1+iBcAANyWriswvfDCC3ruuefUpk0b+fv765NPPpExRu+///7Nqg8AAMDlruuS3Icffqg333xTy5Yt0zfffKPvv/9eH3/8sVJSUm5WfQAAAC53XYHp2LFj6tSpk/15mzZtZLPZdPLkyVwvDAAAwF1cV2BKSkqSn5+fwzBvb29dvXo1V4sCAABwJ9fVh8kYo4EDB8rX19c+7MqVK3rkkUeUN29e+7Cvv/46W8v75Zdf9PLLL2vbtm06deqUFi9erB49ejisb9KkSXrnnXcUExOjxo0ba968eXQuBwAAt9R1nWGKiIhQcHCwAgMD7Y/+/fsrNDTUYVh2Xbp0STVr1tTcuXMzHD9jxgy9/vrrmj9/vjZt2qS8efOqffv2unLlyvWUDQAAcEOu6wzTggULcnXlHTt2VMeOHTMcZ4zRrFmzNGHCBHXv3l3StU7nRYsW1TfffKM+ffrkai0AAACZydGNK2+Fw4cP6/Tp02rTpo19WGBgoOrXr68NGzZkOl9CQoJiY2MdHgAAADfCbQPT6dOnJUlFixZ1GF60aFH7uIxMmzbN4fJgyZIlb2qdAADg9ue2gSmnxo8frwsXLtgfx48fd3VJAADgX85tA1NISIgk6cyZMw7Dz5w5Yx+XEV9fX+XPn9/hAQAAcCPcNjCVLl1aISEhWrVqlX1YbGysNm3apIYNG7qwMgAAcKe5rm/J5ba4uDgdPHjQ/vzw4cOKjIxUoUKFFBYWppEjR+r5559X+fLlVbp0aU2cOFGhoaEO92oCAAC42VwamLZu3aqWLVvan48aNUrStfs9LVy4UE8//bQuXbqkhx56SDExMWrSpImWLl3qdLdxAACAm8mlgalFixYyxmQ63mazacqUKZoyZcotrAoAAMCR2/ZhAgAAcBcEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsu/fFdwJ1ERUW5ugQFBQUpLCzM1WUAANIhMOGOF/f3GclmU//+/V1divzz5NHeqChCEwC4GQIT7nhXLsZIxqjtlDkqUaWmy+qIPrRP348dqujoaAITALgZAhPwfwqGl1dIlbtcXQYAwA3R6RsAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMCCl6sLAOAoKirK1SUoKChIYWFhri4DANwGgQlwE3F/n5FsNvXv39/Vpcg/Tx7tjYoiNAHA/yEwAW7iysUYyRi1nTJHJarUdFkd0Yf26fuxQxUdHU1gAoD/Q2AC3EzB8PIKqXKXq8sAAKRBp28AAAALBCYAAAALBCYAAAALBCYAAAALBCYAAAALBCYAAAALBCYAAAALBCYAAAALBCYAAAALBCYAAAALbh2YnnvuOdlsNodHpUqVXF0WAAC4w7j9b8lVrVpVK1eutD/38nL7kgEAwG3G7dOHl5eXQkJCsj19QkKCEhIS7M9jY2NvRlkAAOAO4vaB6cCBAwoNDZWfn58aNmyoadOmKSwsLNPpp02bpsmTJ9/CCoHbU1RUlEvXHxQUlOV7HQBuJbcOTPXr19fChQtVsWJFnTp1SpMnT1bTpk21a9cuBQQEZDjP+PHjNWrUKPvz2NhYlSxZ8laVDPzrxf19RrLZ1L9/f5fW4Z8nj/ZGRRGaALgFtw5MHTt2tP9do0YN1a9fX6VKldLnn3+uIUOGZDiPr6+vfH19b1WJwG3nysUYyRi1nTJHJarUdEkN0Yf26fuxQxUdHU1gAuAW3DowpVegQAFVqFBBBw8edHUpwG2vYHh5hVS5y9VlAIBbcOvbCqQXFxenQ4cOqVixYq4uBQAA3EHcOjA99dRTWrt2rY4cOaLffvtN99xzjzw9PdW3b19XlwYAAO4gbn1J7q+//lLfvn117tw5FSlSRE2aNNHGjRtVpEgRV5cGAADuIG4dmBYtWuTqEgAAANz7khwAAIA7IDABAABYIDABAABYIDABAABYIDABAABYIDABAABYIDABAABYIDABAABYIDABAABYIDABAABYIDABAABYIDABAABYcOsf3wVwZ4uKinJ1CQoKClJYWJirywDgYgQmAG4n7u8zks2m/v37u7oU+efJo71RUYQm4A5HYALgdq5cjJGMUdspc1SiSk2X1RF9aJ++HztU0dHRBCbgDkdgAuC2CoaXV0iVu1xdBgDQ6RsAAMAKgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMACgQkAAMAC92ECAAv8RAsAAhMAZIKfaAGQisAEAJngJ1oApCIwAYAFfqIFAJ2+AQAALBCYAAAALBCYAAAALBCYAAAALBCYAAAALBCYAAAALHBbAQD4l3D1Hce52zjuZAQmAHBz7nLHce42jjsZgQkA3Jw73HGcu43jTkdgAoB/Ce44DrgOnb4BAAAscIYJAJBtru54LtH5HK5BYAIAWHKXjucSnc/hGgQmAIAld+h4LtH5HK5DYAIAZBsdz3GnotM3AACABc4wAQD+deh8jluNwAQA+Neg8zlchcAEAPjXoPM5XIXABAD413GXzueuvjTIZcFbh8AEAMB1cpdLg1wWvHUITAAAXCd3uDTIZcFbi8AEAEAOuculQdx8BCYAAP7FXN2PSroz+lIRmAAA+Bdyl35U0p3Rl4rABADAv5A79KOS7py+VAQmAAD+xehHdWvwW3IAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAWCEwAAAAW/hWBae7cuQoPD5efn5/q16+vzZs3u7okAABwB3H7wPTZZ59p1KhRmjRpkrZv366aNWuqffv2Onv2rKtLAwAAdwi3D0wzZ87Ugw8+qEGDBqlKlSqaP3++8uTJo/fff9/VpQEAgDuEl6sLyEpiYqK2bdum8ePH24d5eHioTZs22rBhQ4bzJCQkKCEhwf78woULkqTY2NibW2w2xcXFSZJO7Y5UYvwll9URfWi/JOnvvX/IW+aOrsMdaqAO96uBOtyvBupwvxok6dzhA5Kufb7l5mdt6rKMcd1rc2Dc2IkTJ4wk89tvvzkMHzNmjKlXr16G80yaNMlI4sGDBw8ePHjcBo/jx4/fishhya3PMOXE+PHjNWrUKPvzmJgYlSpVSseOHVNgYKALK7smNjZWJUuW1PHjx5U/f35XlyPJ/Wpyt3ok96vJ3eqR3K8md6tHcr+a3K0eyf1qcrd6JPer6WbVY4zRxYsXFRoammvLvBFuHZiCgoLk6empM2fOOAw/c+aMQkJCMpzH19dXvr6+TsMDAwPd4sBKlT9/freqR3K/mtytHsn9anK3eiT3q8nd6pHcryZ3q0dyv5rcrR7J/Wq6GfW4w4mOVG7d6dvHx0e1a9fWqlWr7MNSUlK0atUqNWzY0IWVAQCAO4lbn2GSpFGjRikiIkJ16tRRvXr1NGvWLF26dEmDBg1ydWkAAOAO4faBqXfv3vr777/17LPP6vTp07rrrru0dOlSFS1aNFvz+/r6atKkSRlepnMFd6tHcr+a3K0eyf1qcrd6JPeryd3qkdyvJnerR3K/mtytHsn9anK3em4WmzHu8n09AAAA9+TWfZgAAADcAYEJAADAAoEJAADAAoEJAADAwm0dmObOnavw8HD5+fmpfv362rx5s8tq+eWXX9S1a1eFhobKZrPpm2++cVktkjRt2jTVrVtXAQEBCg4OVo8ePbRv3z6X1jRv3jzVqFHDfvOzhg0basmSJS6tKa2XXnpJNptNI0eOdFkNzz33nGw2m8OjUqVKLqtHkk6cOKH+/furcOHC8vf3V/Xq1bV161aX1RMeHu60jWw2m4YNG+aSepKTkzVx4kSVLl1a/v7+Klu2rKZOnery38e6ePGiRo4cqVKlSsnf31+NGjXSli1bbsm6rdpDY4yeffZZFStWTP7+/mrTpo0OHDjg0pq+/vprtWvXToULF5bNZlNkZKTL6rl69arGjh2r6tWrK2/evAoNDdUDDzygkydPuqwm6Vr7VKlSJeXNm1cFCxZUmzZttGnTppta06102wamzz77TKNGjdKkSZO0fft21axZU+3bt9fZs2ddUs+lS5dUs2ZNzZ071yXrT2/t2rUaNmyYNm7cqBUrVujq1atq166dLl1y3Q8ClyhRQi+99JK2bdumrVu3qlWrVurevbt2797tsppSbdmyRW+99ZZq1Kjh6lJUtWpVnTp1yv5Yv369y2r5559/1LhxY3l7e2vJkiXas2ePXn31VRUsWNBlNW3ZssVh+6xYsUKSdN9997mknunTp2vevHmaM2eOoqKiNH36dM2YMUNvvPGGS+pJNXToUK1YsUIfffSRdu7cqXbt2qlNmzY6ceLETV+3VXs4Y8YMvf7665o/f742bdqkvHnzqn379rpy5YrLarp06ZKaNGmi6dOn37QasltPfHy8tm/frokTJ2r79u36+uuvtW/fPnXr1s1lNUlShQoVNGfOHO3cuVPr169XeHi42rVrp7///vum1nXLuPSX7G6ievXqmWHDhtmfJycnm9DQUDNt2jQXVnWNJLN48WJXl+Hg7NmzRpJZu3atq0txULBgQfPuu++6tIaLFy+a8uXLmxUrVpjmzZubESNGuKyWSZMmmZo1a7ps/emNHTvWNGnSxNVlZGnEiBGmbNmyJiUlxSXr79y5sxk8eLDDsJ49e5p+/fq5pB5jjImPjzeenp7mhx9+cBh+9913m2eeeeaW1pK+PUxJSTEhISHm5Zdftg+LiYkxvr6+5tNPP3VJTWkdPnzYSDI7duy4JbVY1ZNq8+bNRpI5evSo29R04cIFI8msXLnyltR0s92WZ5gSExO1bds2tWnTxj7Mw8NDbdq00YYNG1xYmfu6cOGCJKlQoUIuruSa5ORkLVq0SJcuXXL5z+AMGzZMnTt3djieXOnAgQMKDQ1VmTJl1K9fPx07dsxltXz33XeqU6eO7rvvPgUHB6tWrVp65513XFZPeomJifrf//6nwYMHy2azuaSGRo0aadWqVdq/f78k6ffff9f69evVsWNHl9QjSUlJSUpOTpafn5/DcH9/f5eesZSkw4cP6/Tp0w7vt8DAQNWvX5/2OwsXLlyQzWZTgQIFXF2KpGvvvbfffluBgYGqWbOmq8vJFW5/p++ciI6OVnJystPdwIsWLaq9e/e6qCr3lZKSopEjR6px48aqVq2aS2vZuXOnGjZsqCtXrihfvnxavHixqlSp4rJ6Fi1apO3bt9+yvh1W6tevr4ULF6pixYo6deqUJk+erKZNm2rXrl0KCAi45fX8+eefmjdvnkaNGqX//ve/2rJli5544gn5+PgoIiLilteT3jfffKOYmBgNHDjQZTWMGzdOsbGxqlSpkjw9PZWcnKwXXnhB/fr1c1lNAQEBatiwoaZOnarKlSuraNGi+vTTT7VhwwaVK1fOZXVJ0unTpyUpw/Y7dRwcXblyRWPHjlXfvn1d/mO8P/zwg/r06aP4+HgVK1ZMK1asUFBQkEtryi23ZWDC9Rk2bJh27drl8v8sJalixYqKjIzUhQsX9OWXXyoiIkJr1651SWg6fvy4RowYoRUrVjj9J+4qac9K1KhRQ/Xr11epUqX0+eefa8iQIbe8npSUFNWpU0cvvviiJKlWrVratWuX5s+f7xaB6b333lPHjh0VGhrqsho+//xzffzxx/rkk09UtWpVRUZGauTIkQoNDXXpNvroo480ePBgFS9eXJ6enrr77rvVt29fbdu2zWU14fpdvXpV999/v4wxmjdvnqvLUcuWLRUZGano6Gi98847uv/++7Vp0yYFBwe7urQbdltekgsKCpKnp6fOnDnjMPzMmTMKCQlxUVXuafjw4frhhx+0evVqlShRwtXlyMfHR+XKlVPt2rU1bdo01axZU7Nnz3ZJLdu2bdPZs2d19913y8vLS15eXlq7dq1ef/11eXl5KTk52SV1pVWgQAFVqFBBBw8edMn6ixUr5hRmK1eu7NLLhKmOHj2qlStXaujQoS6tY8yYMRo3bpz69Omj6tWra8CAAXryySc1bdo0l9ZVtmxZrV27VnFxcTp+/Lg2b96sq1evqkyZMi6tK7WNpv22lhqWjh49qhUrVrj87JIk5c2bV+XKlVODBg303nvvycvLS++9956ry8oVt2Vg8vHxUe3atbVq1Sr7sJSUFK1atcrl/WHchTFGw4cP1+LFi/Xzzz+rdOnSri4pQykpKUpISHDJulu3bq2dO3cqMjLS/qhTp4769eunyMhIeXp6uqSutOLi4nTo0CEVK1bMJetv3Lix0+0o9u/fr1KlSrmknrQWLFig4OBgde7c2aV1xMfHy8PDsan19PRUSkqKiypylDdvXhUrVkz//POPli1bpu7du7u0ntKlSyskJMSh/Y6NjdWmTZtov9NIDUsHDhzQypUrVbhwYVeXlCFXtuG57ba9JDdq1ChFRESoTp06qlevnmbNmqVLly5p0KBBLqknLi7O4SzA4cOHFRkZqUKFCiksLOyW1zNs2DB98skn+vbbbxUQEGDvGxAYGCh/f/9bXo8kjR8/Xh07dlRYWJguXryoTz75RGvWrNGyZctcUk9AQIBTn668efOqcOHCLuvr9dRTT6lr164qVaqUTp48qUmTJsnT01N9+/Z1ST1PPvmkGjVqpBdffFH333+/Nm/erLfffltvv/22S+pJlZKSogULFigiIkJeXq5t5rp27aoXXnhBYWFhqlq1qnbs2KGZM2dq8ODBLq1r2bJlMsaoYsWKOnjwoMaMGaNKlSrdkjbSqj0cOXKknn/+eZUvX16lS5fWxIkTFRoaqh49erispvPnz+vYsWP2ex2l/qMQEhJyU858ZVVPsWLFdO+992r79u364YcflJycbG/DCxUqJB8fn1yvx6qmwoUL64UXXlC3bt1UrFgxRUdHa+7cuTpx4oTLbumR61z8Lb2b6o033jBhYWHGx8fH1KtXz2zcuNFltaxevdpIcnpERES4pJ6MapFkFixY4JJ6jDFm8ODBplSpUsbHx8cUKVLEtG7d2ixfvtxl9WTE1bcV6N27tylWrJjx8fExxYsXN7179zYHDx50WT3GGPP999+batWqGV9fX1OpUiXz9ttvu7QeY4xZtmyZkWT27dvn6lJMbGysGTFihAkLCzN+fn6mTJky5plnnjEJCQkureuzzz4zZcqUMT4+PiYkJMQMGzbMxMTE3JJ1W7WHKSkpZuLEiaZo0aLG19fXtG7d+qbvS6uaFixYkOH4SZMm3fJ6Um9tkNFj9erVN6Ueq5ouX75s7rnnHhMaGmp8fHxMsWLFTLdu3czmzZtvWj23ms0YF99uFgAAwM3dln2YAAAAchOBCQAAwAKBCQAAwAKBCQAAwAKBCQAAwAKBCQAAwAKBCQAAwAKBCQAAwAKBCcBtY+HChSpQoICrywBwGyIwAbgpWrRooZEjR7q6DADIFQQmAAAACwQmALlu4MCBWrt2rWbPni2bzSabzaYjR45o7dq1qlevnnx9fVWsWDGNGzdOSUlJ9vnCw8M1a9Ysh2Xdddddeu655+zPY2Ji9PDDD6to0aLy8/NTtWrV9MMPPzjMs2zZMlWuXFn58uVThw4ddOrUKYfx7777ripXriw/Pz9VqlRJb775pn1cYmKihg8frmLFisnPz0+lSpXStGnTcm/jAPhX8nJ1AQBuP7Nnz9b+/ftVrVo1TZkyRZKUnJysTp06aeDAgfrwww+1d+9ePfjgg/Lz83MIRFlJSUlRx44ddfHiRf3vf/9T2bJltWfPHnl6etqniY+P1yuvvKKPPvpIHh4e6t+/v5566il9/PHHkqSPP/5Yzz77rObMmaNatWppx44devDBB5U3b15FRETo9ddf13fffafPP/9cYWFhOn78uI4fP57r2wjAvwuBCUCuCwwMlI+Pj/LkyaOQkBBJ0jPPPKOSJUtqzpw5stlsqlSpkk6ePKmxY8fq2WeflYeH9QnvlStXavPmzYqKilKFChUkSWXKlHGY5urVq5o/f77Kli0rSRo+fLg9tEnSpEmT9Oqrr6pnz56SpNKlS2vPnj166623FBERoWPHjql8+fJq0qSJbDabSpUqlSvbBMC/G5fkANwSUVFRatiwoWw2m31Y48aNFRcXp7/++itby4iMjFSJEiXsYSkjefLksYclSSpWrJjOnj0rSbp06ZIOHTqkIUOGKF++fPbH888/r0OHDkm6djkxMjJSFStW1BNPPKHly5fn5OUCuM1whgmA2/Dw8JAxxmHY1atX7X/7+/tbLsPb29vhuc1msy8zLi5OkvTOO++ofv36DtOlXta7++67dfjwYS1ZskQrV67U/fffrzZt2ujLL7+8/hcE4LZBYAJwU/j4+Cg5Odn+vHLlyvrqq69kjLGfZfr1118VEBCgEiVKSJKKFCni0EE7NjZWhw8ftj+vUaOG/vrrL+3fvz/Ls0yZKVq0qEJDQ/Xnn3+qX79+mU6XP39+9e7dW71799a9996rDh066Pz58ypUqNB1rxPA7YHABOCmCA8P16ZNm3TkyBHly5dPjz32mGbNmqXHH39cw4cP1759+zRp0iSNGjXK3n+pVatWWrhwobp27aoCBQro2WefdejQ3bx5czVr1ky9evXSzJkzVa5cOe3du1c2m00dOnTIVl2TJ0/WE088ocDAQHXo0EEJCQnaunWr/vnnH40aNUozZ85UsWLFVKtWLXl4eOiLL75QSEgIN8QE7nD0YQJwUzz11FPy9PRUlSpVVKRIEV29elU//fSTNm/erJo1a+qRRx7RkCFDNGHCBPs848ePV/PmzdWlSxd17txZPXr0cOiPJElfffWV6tatq759+6pKlSp6+umnHc5kWRk6dKjeffddLViwQNWrV1fz5s21cOFClS5dWpIUEBCgGTNmqE6dOqpbt66OHDmin376KVud0gHcvmwmfYcBAAAAOOBfJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAsEJgAAAAv/D5CiNDiSUtJVAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "g = sns.histplot(data = by_second, x=by_second.touches, stat='percent', color='#4CC9F0', bins=13)\n",
    "g.text(x=0.5, y=1.1, s='Smartphone screen interaction per second:', fontsize=16, weight='bold', ha='center', va='bottom', transform=g.transAxes)\n",
    "g.text(x=0.5, y=1.05, s='Distribution of touches per second, obtained by grouping every touch event.', fontsize=8, alpha=0.75, ha='center', va='bottom', transform=g.transAxes)\n",
    "g.set_xticks(range(14))\n",
    "g.set_xticklabels([0,1,2,3,4,5,6,7,8,9,10,11,12,13])\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {
    "image-2.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "7efcffd1",
   "metadata": {},
   "source": [
    "## Operations in the next sections:\n",
    "- Cleaning data from lacking / dropout users\n",
    "- Applying the concept of circadian time on days and hours\n",
    "\n",
    "Very obviously, only the users that gave enough days to be useful the experiment should be considered. Below I plotted separately the distribution of events (seconds in which there is at least one touch) over the days in which data was collected. \n",
    "\n",
    "Ideally, the observation should the 15 days of maximum observation, **starting from when most users logged** and **ending when they start to drop out massively**. Personally, I choose the two week period starting from 2020/11/12 and ending 2020/11/27, which is also the same time around the WeNet project sent a notification each half-hour of the day to users - in the time diary data of the project. It will help with more accuracy for activities and locations that the user was in while using the smartphone.\n",
    "### Distribution: count of events (grouped by second) registered between the start and end of observation period\n",
    "![image-2.png](attachment:image-2.png)\n",
    "\n",
    "What also needs to be done is to mark each row as belonging to a specific circadian day - differently from clock time, but not independently. \n",
    "The logic is exemplified as follows: when one user is scrolling their phone from 23:59:00 until 00:20:10 and then went to sleep and did not wake up until the morning after, and we want to group per day (for visualization and analysis purposes) the times they were scrolling, we should count everything that happened from 00:00:00 to 00:20:10 and consider it part of the previous day. This logic should seamlessy apply to the many students that don't stop their day before 24:00:00.  \n",
    "\n",
    "According to prof. Bison and the WeNet team, an ideal breakpoint for counting circadian day activities is between 04:30:00 and 05:00:00, because at that time everyone usually has finished their day and the next will start in a few minutes or hours. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "9b4066b6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 11 s, sys: 1.84 s, total: 12.8 s\n",
      "Wall time: 12.8 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "by_second = pd.read_csv('Processed Data/by_second_clean.csv')\n",
    "timestamp, day = pd.to_datetime(by_second.timestamp, yearfirst=True), pd.to_datetime(by_second.day, yearfirst=True)\n",
    "by_second.timestamp, by_second.day = timestamp, day\n",
    "by_second = by_second.drop('Unnamed: 0',axis='columns')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "e1d79db2-e53c-45e0-9443-5f47e1d9290e",
   "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>userid</th>\n",
       "      <th>day</th>\n",
       "      <th>timestamp</th>\n",
       "      <th>touches</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:00</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:03</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:06</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:53</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-22</td>\n",
       "      <td>2020-11-22 00:00:54</td>\n",
       "      <td>1.0</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>19859169</th>\n",
       "      <td>208</td>\n",
       "      <td>2020-11-24</td>\n",
       "      <td>2020-11-24 23:06:12</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19859170</th>\n",
       "      <td>208</td>\n",
       "      <td>2020-11-24</td>\n",
       "      <td>2020-11-24 23:06:13</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19859171</th>\n",
       "      <td>208</td>\n",
       "      <td>2020-11-24</td>\n",
       "      <td>2020-11-24 23:06:14</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19859172</th>\n",
       "      <td>208</td>\n",
       "      <td>2020-11-24</td>\n",
       "      <td>2020-11-24 23:06:18</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19859173</th>\n",
       "      <td>208</td>\n",
       "      <td>2020-11-24</td>\n",
       "      <td>2020-11-24 23:06:19</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>19859174 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          userid        day           timestamp  touches\n",
       "0              1 2020-11-22 2020-11-22 00:00:00      6.0\n",
       "1              1 2020-11-22 2020-11-22 00:00:03      1.0\n",
       "2              1 2020-11-22 2020-11-22 00:00:06      2.0\n",
       "3              1 2020-11-22 2020-11-22 00:00:53      2.0\n",
       "4              1 2020-11-22 2020-11-22 00:00:54      1.0\n",
       "...          ...        ...                 ...      ...\n",
       "19859169     208 2020-11-24 2020-11-24 23:06:12      2.0\n",
       "19859170     208 2020-11-24 2020-11-24 23:06:13      1.0\n",
       "19859171     208 2020-11-24 2020-11-24 23:06:14      2.0\n",
       "19859172     208 2020-11-24 2020-11-24 23:06:18      1.0\n",
       "19859173     208 2020-11-24 2020-11-24 23:06:19      2.0\n",
       "\n",
       "[19859174 rows x 4 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "by_second"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c6fd1cf8",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████| 201/201 [00:02<00:00, 82.06it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "N: 201\n",
      "Done!\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "IDs=by_second.userid.unique()\n",
    "IDs.sort()\n",
    "\n",
    "#data structure:\n",
    "'''\n",
    "{id: [2020-11-12, 2020-11,13, 2020-11-14,...]}   #for every ID, there is a list of days in which there is at least one observation from them. \n",
    "'''\n",
    "user_activity = dict()\n",
    "for ID in tqdm(IDs):\n",
    "    id_act = by_second[by_second.userid == ID]\n",
    "    # adding to the dictionary\n",
    "    user_activity[f'{ID}'] = pd.Series(id_act.day.unique()).sort_values()\n",
    "\n",
    "print('N:',len(user_activity.keys())) #should be 201\n",
    "print(\"Done!\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f37a41b3",
   "metadata": {},
   "source": [
    "Now, I need to know who went through the whole period and who didn't. This means:\n",
    "- counting the days the user sent "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a0823a2d",
   "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>userid</th>\n",
       "      <th>start</th>\n",
       "      <th>finish</th>\n",
       "      <th>tot</th>\n",
       "      <th>cons</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>2020-11-12</td>\n",
       "      <td>2020-11-25</td>\n",
       "      <td>13</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   userid       start      finish  tot   cons\n",
       "0       1  2020-11-12  2020-11-25   13  False"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def gantt_data_org(k, v):\n",
    "    #gantt data is named after the Gantt chart, although it's not present here\n",
    "    \n",
    "    #S is the Series of days; it's best if it's a consecutive 14 days (regardless of when they start)\n",
    "    \n",
    "    tot = 0\n",
    "    start = np.nan\n",
    "    end = np.nan\n",
    "    consecutive = False\n",
    "    \n",
    "    if len(v):  #if it's not empty, it will check for continuity\n",
    "        tot = (pd.to_datetime(v[-1]) - pd.to_datetime(v[0])).days\n",
    "        start, end = v[0], v[-1]\n",
    "        for i in range(len(v)-1):\n",
    "            v[i]  #current\n",
    "            v[i+1] #next\n",
    "            difference = (pd.to_datetime(v[i+1]) - pd.to_datetime(v[i])).days\n",
    "            if difference >= 2:    # if there is more than one day of difference it means that there was one \n",
    "                                   # day in the data with no observations\n",
    "                consecutive = False\n",
    "                break\n",
    "            else:\n",
    "                consecutive = True\n",
    "                \n",
    "    return pd.DataFrame([[k, start, end, tot, consecutive]], \n",
    "                        columns=['userid', 'start', 'finish', 'tot', 'cons'])\n",
    "\n",
    "#example of output:\n",
    "gantt_data_org(1, [\"2020-11-12\", \"2020-11-13\", \"2020-11-14\", \"2020-11-15\", \"2020-11-16\", \n",
    "                   \"2020-11-17\",   \"2020-11-18\", \"2020-11-19\", \"2020-11-25\"])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a4a8c2e2-7d8f-4943-8ce6-831f203111bc",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "41dde2d3",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_29266/81862931.py:4: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation.\n",
      "  gantt_data = pd.concat([gantt_data, res], ignore_index=True)\n"
     ]
    }
   ],
   "source": [
    "gantt_data = pd.DataFrame(columns=[\"userid\", \"start\", \"finish\", \"tot\", \"cons\"])\n",
    "for ID, days in user_activity.items():\n",
    "    res = gantt_data_org(ID, days.tolist())\n",
    "    gantt_data = pd.concat([gantt_data, res], ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "332d7911",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tot\n",
       "0     10\n",
       "1      2\n",
       "2      2\n",
       "3      4\n",
       "4      1\n",
       "5      3\n",
       "6      2\n",
       "7      2\n",
       "8      1\n",
       "9      4\n",
       "10     1\n",
       "11     3\n",
       "12     1\n",
       "13     3\n",
       "14     4\n",
       "15     3\n",
       "17     2\n",
       "18     3\n",
       "19     2\n",
       "20     7\n",
       "22     1\n",
       "23     5\n",
       "24     1\n",
       "26     3\n",
       "27     1\n",
       "28     5\n",
       "29    13\n",
       "30    49\n",
       "31    34\n",
       "32    14\n",
       "33     5\n",
       "34     4\n",
       "35     5\n",
       "36     1\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gantt_data.tot.value_counts().sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "c36d2195",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Users with 14 days or less of recorded activity:\n",
      "\u001b[1m43\u001b[0m\n",
      "15 days or more of recorded activity: \n",
      "\u001b[1m158\u001b[0m\n",
      "\n",
      "Total: \n",
      "\u001b[1m201\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "less_than_14 = sum(gantt_data.tot.value_counts().sort_index()[0:15])\n",
    "more_than_14 = sum(gantt_data.tot.value_counts().sort_index()[15:])\n",
    "total = less_than_14 + more_than_14\n",
    "print(f\"Users with 14 days or less of recorded activity:\\n{color.BOLD}{less_than_14}{color.END}\")\n",
    "print(f\"15 days or more of recorded activity: \\n{color.BOLD}{more_than_14}{color.END}\")\n",
    "print()\n",
    "print(f\"Total: \\n{color.BOLD}{total}{color.END}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "40572541",
   "metadata": {},
   "source": [
    "There are 43 users who are absolutely not eligible for my analysis and are considered dropped out because they didn't even register 15 days, but 135 users who did well. Now that the general distribution has been seen, I will apply the same filter (people who did 15 days and people who did 14 or less) but regarding the specific time period of 12th November to the 27th of the same month.\n",
    "\n",
    "As we can see in the figure below, most people dropped out around the 30th day or so because they complied with a continuation of the experiment, which was facultative for extending the data. Unfortunately, since in another dataset containing the time diary data, in the 12-27 period of November there is supposedly an answer at least every 30 minutes, and after there is an answer at least every 60 minutes (the notifications from the app recording the data were fired at different rates from one period to the other), making it less accurate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "608a9b21",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fluve/anaconda3/lib/python3.11/site-packages/seaborn/_oldcore.py:1119: FutureWarning: use_inf_as_na option is deprecated and will be removed in a future version. Convert inf values to NaN before operating instead.\n",
      "  with pd.option_context('mode.use_inf_as_na', True):\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "g = gantt_data[gantt_data.cons==True]\n",
    "\n",
    "sns.histplot(data=g, x=g.tot, stat=\"percent\", color=\"#660033\")\n",
    "plt.suptitle(\"Each bar height represents the frequency of days for users from start until the end\", y=0.91, ha='center', fontsize=8)\n",
    "plt.title(\"Consecutive days for users before dropping out\", y=1.04,fontsize=18)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "6b396b56",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████| 201/201 [00:03<00:00, 55.96it/s]\n",
      "/tmp/ipykernel_29266/2556236612.py:16: FutureWarning: The behavior of DataFrame concatenation with empty or all-NA entries is deprecated. In a future version, this will no longer exclude empty or all-NA columns when determining the result dtypes. To retain the old behavior, exclude the relevant entries before the concat operation.\n",
      "  gantt_data = pd.concat([gantt_data, res], ignore_index=True)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "N: 201\n",
      "New Filtering done!\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tot\n",
       "0      11\n",
       "1       3\n",
       "2       4\n",
       "3       3\n",
       "4       3\n",
       "5       3\n",
       "7       3\n",
       "8       1\n",
       "9       2\n",
       "10      2\n",
       "11      2\n",
       "12      4\n",
       "13      9\n",
       "14      4\n",
       "15    147\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Filtering for period of interest\n",
    "user_activity = dict()\n",
    "for ID in tqdm(IDs):\n",
    "    id_act = by_second[by_second.userid == ID]\n",
    "    id_act = id_act[id_act.day.dt.month == 11]\n",
    "    id_act = id_act[id_act.day.dt.day <= 27]\n",
    "    id_act = id_act[id_act.day.dt.day >= 12]\n",
    "    # adding to the dictionary\n",
    "    user_activity[f'{ID}'] = pd.Series(id_act.day.unique()).sort_values().reset_index()\n",
    "print(\"N:\",len(user_activity.keys()))\n",
    "print(\"New Filtering done!\")\n",
    "\n",
    "gantt_data = pd.DataFrame(columns=[\"userid\", \"start\", \"finish\", \"tot\", \"cons\"])\n",
    "for ID, days in user_activity.items():\n",
    "    res = gantt_data_org(ID, days[0].tolist())\n",
    "    gantt_data = pd.concat([gantt_data, res], ignore_index=True)\n",
    "gantt_data.tot.value_counts().sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "fbec1739",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fluve/anaconda3/lib/python3.11/site-packages/seaborn/_oldcore.py:1119: FutureWarning: use_inf_as_na option is deprecated and will be removed in a future version. Convert inf values to NaN before operating instead.\n",
      "  with pd.option_context('mode.use_inf_as_na', True):\n",
      "/tmp/ipykernel_29266/3114761937.py:7: UserWarning: Tight layout not applied. The bottom and top margins cannot be made large enough to accommodate all axes decorations.\n",
      "  plt.tight_layout()\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 100x100 with 25 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(5, 5, figsize = (1,1))\n",
    "g = gantt_data[gantt_data.cons==True]\n",
    "sns.histplot(data=g, x=g.tot, stat=\"percent\", color=\"#660033\", bins=15)\n",
    "\n",
    "plt.xticks(np.arange(16), list(range(0,16))) \n",
    "plt.suptitle(\"Each bar height represents the frequency of days for users from start until the end\", y=0.86, ha='center', fontsize=8)\n",
    "plt.title(\"Consecutive Days 15 day period version\", y=1.04,fontsize=18)\n",
    "plt.tight_layout()\n",
    "plt.savefig('before_dropout.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8e891272",
   "metadata": {},
   "outputs": [],
   "source": [
    "dropouts = gantt_data[gantt_data.tot <= 14].userid\n",
    "\n",
    "print(\"Expected drop in user id count: from\", gantt_data.shape[0], \"to\", gantt_data.shape[0] - len(dropouts))\n",
    "print(f'Rows of the whole dataset before selection: \\n{color.BOLD}{by_second.shape[0]:,}{color.END}', )\n",
    "df = by_second[~by_second['userid'].isin(dropouts.astype(int))]\n",
    "df = df[df.day.dt.month == 11]\n",
    "df = df[df.day.dt.day <= 28]\n",
    "df = df[df.day.dt.day >= 11]\n",
    "\n",
    "to_keep = df.userid.unique()\n",
    "print(f'Rows of the whole dataset after selection: \\n{color.BOLD}{df.shape[0]:,}{color.END}', )\n",
    "print(\"N° of users:\", len(df.userid.unique()))\n",
    "filename = 'Processed Data/by_second_15.csv'\n",
    "to_keep = df.userid.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "12c9f701",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Saving list of dropouts and the by_second 15 period dataframe\n",
    "df.to_csv(filename)\n",
    "dropouts.to_csv(\"Processed Data/userid_dropped.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "44066edd",
   "metadata": {},
   "source": [
    "# 15 day period - by minute version\n",
    "Now, denoting what time a row belongs to using human day measures requires a certain computational power, so the dataset will be reduced into minutes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e4023eb7",
   "metadata": {},
   "outputs": [],
   "source": [
    "four_million = 4*10**6\n",
    "df_min = pd.DataFrame()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "05aca169",
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = 'Processed Data/by_second_15.csv'\n",
    "for chunk in tqdm(pd.read_csv(filename, chunksize=four_million)): \n",
    "    \n",
    "    temp = chunk[['userid', 'timestamp', 'touches']].copy(deep=True)\n",
    "    temp['timestamp'] = pd.to_datetime(temp['timestamp'], yearfirst=True).copy(deep=True)\n",
    "    temp = temp.groupby(['userid', pd.Grouper(key='timestamp', freq='1Min')])['touches'].sum().reset_index().copy(deep=True)\n",
    "    df_min = pd.concat([df_min, temp], ignore_index=True) \n",
    "    \n",
    "print(\"Done!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d64f343b",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_min.to_csv('Processed Data/by_minute.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9316b817-8dac-441b-b0b5-24755b228876",
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install timedelta -q"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f9c197b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# run once more to avoid duplicates, see issue in file_reduction.ipynb\n",
    "import timedelta\n",
    "\n",
    "df_min = df_min.groupby(['userid', pd.Grouper(key='timestamp', freq='1Min')])['touches'].sum().reset_index().copy(deep=True)\n",
    "df_min['touches'] = df_min.touches.astype(int)\n",
    "print(f'Rows: {df_min.shape[0]:,}')  #should be '682,611'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0e925c8",
   "metadata": {},
   "source": [
    "# Side question:\n",
    "Humans generally stay awake 16 to 18 hours of their 24 hour day, so about 17/24 = 71% is spent awake. Out of these hours, which amount corresponds to the times the phone is checked?\n",
    "Each row corresponds to one minute for one user, so length of the data is \"N rows\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ea6ef884",
   "metadata": {},
   "outputs": [],
   "source": [
    "time = timedelta.Timedelta(df_min.timestamp.max() - df_min.timestamp.min()).total\n",
    "n_users = len(df_min.userid.unique())\n",
    "print(f'{color.BOLD}formula used{color.END}: \\n(N rows) / (total possible minutes awake in the 15 day period)')\n",
    "print()\n",
    "print(f'{df_min.shape[0]:,} / ({time.minutes} x {17/24:.2f} x {n_users}) = {df_min.shape[0] / ((time.minutes)*n_users*(17/24)) * 100:.2f} %')\n",
    "print()\n",
    "print(f\"For all the time the average user is awake,{color.BOLD} the total amount of the minutes is one quarter checking the phone at least once{color.END}.\\nEach minute spent checking the phone has a{color.BOLD} median of {df_min.touches.median():.0f} touches. {color.END}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ca849032",
   "metadata": {},
   "outputs": [],
   "source": [
    "#df_min = df_min.drop(columns=\"day\")\n",
    "days = dict()\n",
    "for i in range(10, 29):\n",
    "    days[i] = i-12 + 1   #days translates to month days to common count of days\n",
    "    \n",
    "def today_or_yesterday(t): \n",
    "    if t.hour < 5:\n",
    "        return days[t.day -1]\n",
    "    else:\n",
    "        return days[t.day]\n",
    "day = [[today_or_yesterday(timestamp) for timestamp in df_min['timestamp']]]\n",
    "df_min.insert(3, 'day', pd.Series(day[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1d641aab",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_min[df_min.userid==1 &\n",
    "       ((df_min['timestamp'].dt.day == 12) & \n",
    "        (df_min['timestamp'].dt.hour >= 23) & (df_min['timestamp'].dt.minute >= 55)| \n",
    "        (df_min['timestamp'].dt.day == 13) & \n",
    "        (df_min['timestamp'].dt.hour <= 9) & (df_min['timestamp'].dt.minute <= 10)) \n",
    "      ].head(25).sort_values(by='timestamp')\n",
    "# Looking at the day indicator below, it stayed the same when needed and changed when needed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "831c0160",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "plt.figure(figsize=(10, 6))\n",
    "g = sns.boxenplot(x='day', y='touches', data=df_min)\n",
    "plt.xlabel('Day')\n",
    "plt.ylabel('Touches')\n",
    "human_days = [' Wed', 'Thu', 'Fri','Sat', 'Sun', 'Mon', 'Tue',] * 2\n",
    "human_days.extend([' Wed', 'Thu', 'Fri', 'Sat', 'Sun'])\n",
    "g.set_xticklabels(human_days)\n",
    "plt.title('Touches Across Days')\n",
    "plt.show()\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "944f2a10",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Graph: I need to make code more readable.\n",
    "\n",
    "out = df_min[df_min.userid.isin([1])]\n",
    "min_range = pd.date_range(start=df_min.timestamp.min(), end=df_min.timestamp.max(), freq='T')\n",
    "complete_data = {'timestamp': min_range}\n",
    "complete_df = pd.DataFrame(complete_data)\n",
    "out = pd.merge(complete_df, pd.DataFrame(out), how='left', on='timestamp')\n",
    "out['touches'] = out['touches'].fillna(0).astype(int)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "42e42d6a",
   "metadata": {},
   "outputs": [],
   "source": [
    "out.timestamp.min()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c5d958ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "out.drop(columns=\"userid\", axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8b3451bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "out['time'] = df_min['timestamp'].dt.hour\n",
    "out = out.groupby(['time', 'day'])['touches'].median().unstack()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1f4844af",
   "metadata": {},
   "outputs": [],
   "source": [
    "g = sns.heatmap(out, cmap=\"flare\")\n",
    "human_days = [' Wed', 'Thu', 'Fri','Sat', 'Sun', 'Mon', 'Tue',] * 2\n",
    "human_days.extend([' Wed', 'Thu', 'Fri'])\n",
    "g.set_xticklabels(human_days)\n",
    "plt.savefig('heatmap_of_use.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "63483970",
   "metadata": {},
   "outputs": [],
   "source": [
    "demographics = pd.read_stata(\"Raw Data/data4diarynew_ITA.dta\")\n",
    "time_diary = pd.read_stata(\"Raw Data/td_ita.dta\")\n",
    "to_keep = pd.read_csv(\"Processed Data/by_minute.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c4566400",
   "metadata": {},
   "source": [
    "Unfortunately, I need to keep only the 147 users that I have sensor data on and for the days that I selected. Out of these, only 2 did not give their data on their social and psychological attributes, so in total there's 145 eligible users. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "c3776fd5-de98-4bac-b42a-2bfbe5f352d4",
   "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>token</th>\n",
       "      <th>w1_idpilot</th>\n",
       "      <th>userid</th>\n",
       "      <th>pilot</th>\n",
       "      <th>w1_A01</th>\n",
       "      <th>nationality</th>\n",
       "      <th>department</th>\n",
       "      <th>Dep_UNITN</th>\n",
       "      <th>cohort</th>\n",
       "      <th>degree</th>\n",
       "      <th>...</th>\n",
       "      <th>Bodykines</th>\n",
       "      <th>Musical</th>\n",
       "      <th>Interpersonal</th>\n",
       "      <th>Intrapersonal</th>\n",
       "      <th>Environmental</th>\n",
       "      <th>Spiritual</th>\n",
       "      <th>w1_A03</th>\n",
       "      <th>w1_A04UNITN</th>\n",
       "      <th>w1_A09UNITN</th>\n",
       "      <th>w1_A10UNITN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>0 rows × 67 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [token, w1_idpilot, userid, pilot, w1_A01, nationality, department, Dep_UNITN, cohort, degree, Extraversion, Agreeableness, Conscientiousness, Neuroticism, Openness, MExtraversion, MAgreeableness, MConscientiousness, MNeuroticism, MOpenness, Pconformity, Ptradition, Pbenov, Punivers, Pself, Pstim, Phedon, Pachieve, Ppower, Psecurity, Popen, Pselfenh, Pselftran, Pconserv, Mconformity, Mtradition, Mbenov, Munivers, Mself, Mstim, Mhedon, Machieve, Mpower, Msecurity, Mopen, Mselfenh, Mselftran, Mconserv, Pexcitements, Psuprapersonal, Pinteractive, Ppromotion, Pexistence, Pnormative, Linguistic, Logicmath, Spatial, Bodykines, Musical, Interpersonal, Intrapersonal, Environmental, Spiritual, w1_A03, w1_A04UNITN, w1_A09UNITN, w1_A10UNITN]\n",
       "Index: []\n",
       "\n",
       "[0 rows x 67 columns]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#I found out about this discrepancy in the descriptive analysis noteboook. \n",
    "# For proof, two users who are present in touch sensor data are 37 and 202.\n",
    "demographics[demographics.userid == 202]\n",
    "demographics[demographics.userid == 37]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ab3fb98b",
   "metadata": {},
   "outputs": [],
   "source": [
    "#users\n",
    "df_td = time_diary[time_diary['id'].isin(to_keep.userid.astype(int))]\n",
    "demographics.userid = demographics.userid.astype(int)\n",
    "df_demographics = demographics[demographics['userid'].isin(df_td.id.unique())]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b4ec27e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "#time\n",
    "df_td = df_td[df_td.datein_answ.dt.month == 11]\n",
    "df_td = df_td[df_td.datein_answ.dt.day <= 28]\n",
    "df_td = df_td[df_td.datein_answ.dt.day >= 11]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "16e321d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_td.datein_answ.max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d1ee3041",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_demographics.to_csv('Processed Data/demographics_15.csv')\n",
    "df_td.to_csv('Processed Data/time_diary_15.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "595da077",
   "metadata": {},
   "source": [
    "# Adding observations with 0 touches\n",
    "In order to visualize things more correctly, I need some structure in the data that also tells the machine that there were no touches in some of the minutes. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6e055e9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "translator = dict()\n",
    "for i in range(9, 29):\n",
    "    translator[i] = i-12 + 1   #translates from dates to chosen count of days\n",
    "\n",
    "def today_or_yesterday(t): \n",
    "    if t.hour < 5:\n",
    "        return translator[t.day - 1]\n",
    "    else:\n",
    "        return translator[t.day]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "65e4e858",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_df = 0\n",
    "data = df_min\n",
    "\n",
    "data['timestamp'] = pd.to_datetime(data['timestamp'])\n",
    "results = pd.DataFrame()\n",
    "for user in tqdm(data.userid.unique()):\n",
    "    subset = data[data.userid==user]\n",
    "    \n",
    "    min_range = pd.date_range(start=\"2020-11-10 00:00:00\", end=\"2020-11-28 23:59:59\", freq='T')\n",
    "    \n",
    "    complete_data = {'timestamp': min_range}\n",
    "    \n",
    "    complete_df = pd.DataFrame(complete_data)\n",
    "    \n",
    "    merged_df = pd.merge(complete_df, pd.DataFrame(subset), how='left', on='timestamp')\n",
    "\n",
    "    # Fill NaN values in 'touches' and 'userid' with 0\n",
    "    merged_df['touches'] = merged_df['touches'].fillna(0).astype(int)\n",
    "    \n",
    "    merged_df['userid'] = merged_df['userid'].fillna(0).astype(int)\n",
    "     \n",
    "    l_userid = [[user for userid in merged_df['userid']]]\n",
    "    merged_df['userid'] = l_userid[0]\n",
    "    \n",
    "    \n",
    "    results = pd.concat([results, merged_df,], ignore_index=True)\n",
    "\n",
    "# Print the result\n",
    "day = [[today_or_yesterday(timestamp) for timestamp in results['timestamp']]]\n",
    "results['day'] = pd.Series(day[0])\n",
    "results[(results.timestamp>=\"2020-11-12 23:50:00\") & (results.userid==2)].head(20).sort_values(by=\"timestamp\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3406ecdf",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(f'Now: the rows are {results.shape[0]:,}. Before: {df_min.shape[0]:,}')\n",
    "print(f'The new dataset is almost {round(results.shape[0]/df_min.shape[0])} times bigger than the original.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c787b4ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "results.to_csv('Processed Data/by_minute_15_with_empty_obs.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5964ab52-bf09-4f15-a98d-db9276464477",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5fd05d8f-4cfe-49e4-bdf4-beceb31702a4",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3099dd46-017b-424c-801b-27e814d325e1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bb28c163-638c-4acb-ba96-511afa6b534b",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}