PyCoffee 2018-02-28 (Trump tweets / Bitcoin price correlation)

pycoffee-2018-02-28.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# pycoffee-2018-02-28 ☕️🐍\n",
    "\n",
    "This time in Porto i/o for a change we decided to get our hands dirty and play with some data visualisation and stuff.\n",
    "\n",
    "We started with a short brainstorm. Its result was an idea to find out if there's any correlation between Donald Trump's tweet sentiments and bitcoin price. At firt sign it may not make any sense, but if you think deeper, all of a sudden there is something inexplicably common in them: they are both unpredictable, they love to climb up and fall down, and they both cause the biggest hype of 2017. \n",
    "\n",
    "Here we go, então!\n",
    "\n",
    "Let's start with Trump tweets. Believe or not, but there's a special website (and a github repository) which automatically downloads each of them and shares them in easy to parse json.zip format. The website is http://www.trumptwitterarchive.com/, and the corresponding repository is this https://github.com/bpb27/trump_tweet_data_archive\n",
    "\n",
    "We downloaded and unzipped the data and then created the data frame."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "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>created_at</th>\n",
       "      <th>favorite_count</th>\n",
       "      <th>id_str</th>\n",
       "      <th>in_reply_to_user_id_str</th>\n",
       "      <th>is_retweet</th>\n",
       "      <th>retweet_count</th>\n",
       "      <th>source</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2016-12-31 18:59:04</td>\n",
       "      <td>0</td>\n",
       "      <td>815271067749060608</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>9529</td>\n",
       "      <td>Twitter for iPhone</td>\n",
       "      <td>RT @realDonaldTrump: Happy Birthday @DonaldJTr...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2016-12-31 18:58:12</td>\n",
       "      <td>55601</td>\n",
       "      <td>815270850916208640</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>9529</td>\n",
       "      <td>Twitter for iPhone</td>\n",
       "      <td>Happy Birthday @DonaldJTrumpJr!\\nhttps://t.co/...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2016-12-31 13:17:21</td>\n",
       "      <td>350860</td>\n",
       "      <td>815185071317676032</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>141853</td>\n",
       "      <td>Twitter for Android</td>\n",
       "      <td>Happy New Year to all, including to my many en...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2016-12-30 22:18:18</td>\n",
       "      <td>84254</td>\n",
       "      <td>814958820980039680</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>23213</td>\n",
       "      <td>Twitter for Android</td>\n",
       "      <td>Russians are playing @CNN and @NBCNews for suc...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2016-12-30 19:46:55</td>\n",
       "      <td>25336</td>\n",
       "      <td>814920722208296960</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>7366</td>\n",
       "      <td>Twitter for iPhone</td>\n",
       "      <td>Join @AmerIcan32, founded by Hall of Fame lege...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           created_at  favorite_count              id_str  \\\n",
       "0 2016-12-31 18:59:04               0  815271067749060608   \n",
       "1 2016-12-31 18:58:12           55601  815270850916208640   \n",
       "2 2016-12-31 13:17:21          350860  815185071317676032   \n",
       "3 2016-12-30 22:18:18           84254  814958820980039680   \n",
       "4 2016-12-30 19:46:55           25336  814920722208296960   \n",
       "\n",
       "   in_reply_to_user_id_str  is_retweet  retweet_count               source  \\\n",
       "0                      NaN        True           9529   Twitter for iPhone   \n",
       "1                      NaN       False           9529   Twitter for iPhone   \n",
       "2                      NaN       False         141853  Twitter for Android   \n",
       "3                      NaN       False          23213  Twitter for Android   \n",
       "4                      NaN       False           7366   Twitter for iPhone   \n",
       "\n",
       "                                                text  \n",
       "0  RT @realDonaldTrump: Happy Birthday @DonaldJTr...  \n",
       "1  Happy Birthday @DonaldJTrumpJr!\\nhttps://t.co/...  \n",
       "2  Happy New Year to all, including to my many en...  \n",
       "3  Russians are playing @CNN and @NBCNews for suc...  \n",
       "4  Join @AmerIcan32, founded by Hall of Fame lege...  "
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df2016 = pd.read_json('condensed_2016.json')\n",
    "df2017 = pd.read_json('condensed_2017.json')\n",
    "df2018 = pd.read_json('condensed_2018.json')\n",
    "\n",
    "trump = pd.concat([df2016, df2017, df2018])\n",
    "trump.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The next step is to estimate the Donald's mood for every tweet. None of us had previous experience with sentiment analysis, so we decided to follow the StackOverflow-driven development model.\n",
    "\n",
    "Quite fast we found a model we want. Naturally, it was non other than venerable [NLTK][]. It has a pre-trained sentiment model [VADER][] which we decided to use without a moment's hesitation.\n",
    "\n",
    "It was much easier than we though. The quality of such a blind application of a foreign model is questionable of course, but don't want to spend way too much time on such minor details as model validation.\n",
    "\n",
    "The sentiment analizer returns takes a phrase and returns an object, consisting of several fields. The \"compound\" attribute of the object is basically a normalized version of the difference between positive and negative elements, found in the text. \n",
    "\n",
    "Looks good, but on top of that we can think of *intensity* of the phrase, which is the sum of `positive` and `negative`\n",
    "\n",
    "[NLTK]: http://www.nltk.org/\n",
    "[VADER]: http://www.nltk.org/howto/sentiment.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from nltk.sentiment.vader import SentimentIntensityAnalyzer\n",
    "sid = SentimentIntensityAnalyzer()\n",
    "\n",
    "def sentiment(row):\n",
    "    score = sid.polarity_scores(row.text)\n",
    "    row['sentiment'] = score['compound']\n",
    "    row['intensity'] = score['pos'] + score['neg']\n",
    "    return row\n",
    "    \n",
    "trump = trump.apply(sentiment, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "That's how the results look like"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>text</th>\n",
       "      <th>sentiment</th>\n",
       "      <th>intensity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>RT @realDonaldTrump: Happy Birthday @DonaldJTrumpJr!\\nhttps://t.co/uRxyCD3hBz</td>\n",
       "      <td>0.6114</td>\n",
       "      <td>0.444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Happy Birthday @DonaldJTrumpJr!\\nhttps://t.co/uRxyCD3hBz</td>\n",
       "      <td>0.6114</td>\n",
       "      <td>0.571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Happy New Year to all, including to my many enemies and those who have fought me and lost so bad...</td>\n",
       "      <td>-0.4911</td>\n",
       "      <td>0.476</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Russians are playing @CNN and @NBCNews for such fools - funny to watch, they don't have a clue! ...</td>\n",
       "      <td>0.2695</td>\n",
       "      <td>0.333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Join @AmerIcan32, founded by Hall of Fame legend @JimBrownNFL32 on 1/19/2017 in Washington, D.C....</td>\n",
       "      <td>0.6249</td>\n",
       "      <td>0.282</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Great move on delay (by V. Putin) - I always knew he was very smart!</td>\n",
       "      <td>0.7257</td>\n",
       "      <td>0.492</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>My Administration will follow two simple rules: https://t.co/ZWk0j4H8Qy</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>'Economists say Trump delivered hope' https://t.co/SjGBgglIuQ</td>\n",
       "      <td>0.4404</td>\n",
       "      <td>0.367</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>not anymore. The beginning of the end was the horrible Iran deal, and now this (U.N.)! Stay stro...</td>\n",
       "      <td>-0.1984</td>\n",
       "      <td>0.251</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>We cannot continue to let Israel be treated with such total disdain and disrespect. They used to...</td>\n",
       "      <td>0.3400</td>\n",
       "      <td>0.398</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                                                                  text  \\\n",
       "0                        RT @realDonaldTrump: Happy Birthday @DonaldJTrumpJr!\\nhttps://t.co/uRxyCD3hBz   \n",
       "1                                             Happy Birthday @DonaldJTrumpJr!\\nhttps://t.co/uRxyCD3hBz   \n",
       "2  Happy New Year to all, including to my many enemies and those who have fought me and lost so bad...   \n",
       "3  Russians are playing @CNN and @NBCNews for such fools - funny to watch, they don't have a clue! ...   \n",
       "4  Join @AmerIcan32, founded by Hall of Fame legend @JimBrownNFL32 on 1/19/2017 in Washington, D.C....   \n",
       "5                                 Great move on delay (by V. Putin) - I always knew he was very smart!   \n",
       "6                              My Administration will follow two simple rules: https://t.co/ZWk0j4H8Qy   \n",
       "7                                        'Economists say Trump delivered hope' https://t.co/SjGBgglIuQ   \n",
       "8  not anymore. The beginning of the end was the horrible Iran deal, and now this (U.N.)! Stay stro...   \n",
       "9  We cannot continue to let Israel be treated with such total disdain and disrespect. They used to...   \n",
       "\n",
       "   sentiment  intensity  \n",
       "0     0.6114      0.444  \n",
       "1     0.6114      0.571  \n",
       "2    -0.4911      0.476  \n",
       "3     0.2695      0.333  \n",
       "4     0.6249      0.282  \n",
       "5     0.7257      0.492  \n",
       "6     0.0000      0.000  \n",
       "7     0.4404      0.367  \n",
       "8    -0.1984      0.251  \n",
       "9     0.3400      0.398  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.set_option('max_colwidth', 100)\n",
    "trump[['text', 'sentiment', 'intensity']].head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looks like it makes some sense finally.\n",
    "\n",
    "To warm up, we decided to display a couple of plots to see if there's anything interesting we can reveal. We use samples to make plots looking less dense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "ax = trump.sample(1000).plot(x='created_at', y='sentiment', style='.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEJCAYAAACaFuz/AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJztvXmcXNV55/19qlsSCAvRllgE2pABBUusEiCMF7DBLyQExiwRYGdCYluxB+Kx/Xpek4yjEMXj8WQ8GZwxEywzDEnGLAbsGDNgbMJijJHQYowlYYFotDQChEQjZLT0Uuf9495TferUuVvVrarb1ef3+eijrqq7PPfc5zznOc8qSik8PDw8PDoXpXYT4OHh4eHRXHhB7+Hh4dHh8ILew8PDo8PhBb2Hh4dHh8MLeg8PD48Ohxf0Hh4eHh0OL+g9PDw8Ohxe0Ht4eHh0OLyg9/Dw8OhweEHv4eHh0eHobjcBAFOnTlWzZ89uNxkeHh4eowpr1qzZqZQ6POm4Qgj62bNns3r16naT4eHh4TGqICJb0hznTTceHh4eHY7Mgl5ELhSRjSKySURucPw+U0QeE5FfishzIvK7+ZDq4eHh4VEPMgl6EekCbgYuAt4LXC0i77UO+wrwPaXUacBVwP/Mg1APDw8Pj/qQ1UZ/JrBJKdULICJ3AZcCG4xjFHBo+PdkYHujRHp4eIx+DA4O0tfXx/79+9tNyqjDQQcdxPTp0xk3blxd52cV9McA24zPfcBZ1jE3Aj8RkT8DDgHOd11IRJYASwBmzpyZkQyP0Yg1W/pZ0buLRXOmsGBWT7vJ8Wgx+vr6mDRpErNnz0ZE2k3OqIFSil27dtHX18exxx5b1zWa4Yy9GrhdKTUd+F3gn0Wk5j5KqeVKqYVKqYWHH54YHeQxyrFmSz8fv3UF/+0nG/n4rStYs6W/3SR5tBj79+9nypQpXshnhIgwZcqUhnZCWQX9K8AM4/P08DsTnwS+B6CUeho4CJhaL4EenYEVvbsYGCpTVjA4VGZF7652k+TRBnghXx8aHbesgn4VcLyIHCsi4wmcrfdbx2wFPhISdyKBoH+jISo9Rj0WzZnC+O4SXQLjukssmjOl3SR5jEG8733vSzzmpptuYu/evbned/v27VxxxRUAPPvsszz44IO5Xj8JmQS9UmoIuB54GHieILpmvYgsE5FLwsP+X+DTIvIr4E7gWuU7kI95LJjVw3c/tYgvfnQu3/3UIm+j92gLfvGLXyQe0wxBf/TRR3PvvfcC7RH0KKXa/m/BggXKw8Ojs7Fhw4bM56ze/Kb61qMvqtWb38yFhkMOOUQppdRjjz2mPvShD6nLL79czZ07V11zzTWqXC6rb37zm2rcuHFq/vz56txzz1VKKfXwww+rRYsWqdNOO01dccUVas+ePUoppWbNmqWWLl2qTjvtNDV//nz1/PPPK6WUevzxx9Upp5yiTjnlFHXqqaeqt99+W7388stq3rx56sCBA2rGjBlq6tSp6pRTTlF33XWXOu6449SOHTuUUkoNDw+r97znPZXPJlzjB6xWKWSsz4z18PAoJJrtwP/lL3/JTTfdxIYNG+jt7eWpp57ic5/7HEcffTSPPfYYjz32GDt37uSrX/0qjzzyCGvXrmXhwoX83d/9XeUaU6dOZe3atXz2s5/lG9/4BgDf+MY3uPnmm3n22Wd58sknOfjggyvHjx8/nmXLlrF48WKeffZZFi9ezCc+8Qm++93vAvDII49wyimnkHeAihf0Hh4ehUSzHfhnnnkm06dPp1Qqceqpp7J58+ZaGlasYMOGDZxzzjmceuqp/OM//iNbtoyUl7nssssAWLBgQeX8c845hy9+8Yv8/d//PW+99Rbd3fFR7H/yJ3/CP/3TPwFw22238cd//Mf5PKCBQhQ18/Dw8LChHfiDQ+WmOPAnTJhQ+burq4uhoaGaY5RSXHDBBdx5552x1zDPv+GGG/i93/s9HnzwQc455xwefvhhDjrooEg6ZsyYwZFHHsmjjz7KM888U9Hu84TX6D08PAqJdjnwJ02axJ49ewBYtGgRTz31FJs2bQLgnXfe4YUXXog9/6WXXuKkk07iy1/+MmeccQa/+c1vIq+v8alPfYpPfOITXHnllXR1deX4NAG8oPfw8CgsFszq4brzjmtplNaSJUu48MILOe+88zj88MO5/fbbufrqqzn55JM5++yzawS3jZtuuon58+dz8sknM27cOC666KKq38877zw2bNjAqaeeyt133w3AJZdcwm9/+9ummG0ARBUg8nHhwoXK16P38OhsPP/885x44ontJqOQWL16NV/4whd48sknI49xjZ+IrFFKLUy6vrfRe3h4eLQRX//61/mHf/iHptjmNbzpxsPDw6ONuOGGG9iyZQvvf//7m3YPL+g9PDw8Ohxe0Ht4eLQMRfAJjkY0Om5e0Ht4eLQEBx10ELt27fLCPiNUWI8+LhY/Cd4Z6+Hh0RJMnz6dvr4+3njDF7PNCt1hql54Qe/h4dESjBs3ru4OSR6NIbPpRkQuFJGNIrJJRG5w/P7fReTZ8N8LIvJWPqR6eHh4eNSDTBq9iHQBNwMXEPSLXSUi9yulKs3BlVJfMI7/M+C0nGj18PDwaAhjtW9xVtPNmcAmpVQvgIjcBVwKbIg4/mrgr+onz8PDwyMf6LLHA0NlxneXxlQDnKymm2OAbcbnvvC7GojILOBY4NH6SPPw8PDID2O5b3EzwyuvAu5VSg27fhSRJSKyWkRWey+8h4dHszGW+xZnNd28AswwPk8Pv3PhKuC6qAsppZYDyyEoapaRDg8PD49M0GWPvY0+GauA40XkWAIBfxVwjX2QiPwO0AM83TCFHh4eHjlhwayeMSXgNTKZbpRSQ8D1wMPA88D3lFLrRWSZiFxiHHoVcJfyKXAeHh4ebUfmhCml1IPAg9Z3S63PNzZGVj4Yq6FUHh4eHiY6NjN2LIdSeXh4eJjo2KJmYzmUysPDw8NExwr6sRxK5eHh4WGiY003YzmUysPDw8NExwp6GLuhVB4eHh4mOtZ04+Hh4eERwAt6Dw8Pjw6HF/QeHh4eHQ4v6AuINVv6ufmxTazZ0t9uUjw8PDoAHe2MHY3wiV4eHh55w2v0BYNP9PIY6/A72vzhNfqCQSd6DQ6VW5bo5WsCeRQFfkfbHHhB3wbECdZWJ3r5ieVRJLh2tJ4fG4cX9C1GGsHaykQvP7E8ioR27GjHAjLb6EXkQhHZKCKbROSGiGP+QEQ2iMh6EbmjcTI7B0WzwY/mmkDeltt50DvaL350rt9d5ohMGr2IdAE3AxcQNAZfJSL3K6U2GMccD/w5cI5Sql9EjsiT4NGOomkso7UmkDc5dS586ZL8kdV0cyawSSnVCyAidwGXAhuMYz4N3KyU6gdQSu3Ig9AsKLJzsYiC1ZxYRR47E97k5OGRHlkF/THANuNzH3CWdcwJACLyFNAF3KiU+rF9IRFZAiwBmDlzZkYyotFMTS8vIZhFY2ml4B1NWnLRdkadgNGyyOeNsfDczXDGdgPHA+cC04GfichJSqm3zIOUUsuB5QALFy7MrbdsszS9dgjBVt9zNGnJRdwZjWbcsXIrS3+4jrJShV/k88RoUm4aQVZn7CvADOPz9PA7E33A/UqpQaXUy8ALBIK/JWiWc7GVTlTtZPz+2r6WOm5Hm2N2wawerjvvuI6cmK3Emi39LP3hOobKirKCgQIECbQKRQuOaBayavSrgONF5FgCAX8VcI11zL8AVwP/W0SmEphyehslNC2apem1ylRgahjdJaG7q8TwcGvME/WO3VjY+nYyVvTuYrg8sqkuiTSVv4vEK2PFBJhJ0CulhkTkeuBhAvv7bUqp9SKyDFitlLo//O2jIrIBGAb+g1KqpctkM7z2rTIVmBrGcFmx+MwZHHPYwS2bGFnHbqxsfTsZi+ZMYcK4EgODZUolYdml85vyDovIK2PFBJjZRq+UehB40PpuqfG3Ar4Y/usotCLsy9YwLj99eqGZbzTZ9T3caIcSUyReGQvhnD4ztomoZ5s62jSMere+RdvCj3W0Q4kZDWaSTuFTCRTw9mLhwoVq9erV7SYjV9jb1KUXz6N/70BhGCZPBs56rSJu4fNEpwiHZmA0jc1o4FMRWaOUWph0nNfomwRzmzowWM4UupZ2MtR7XN4MnFUbLOoWPg+MBuHQTowmM0kn8emYFvTN1C7MbaqIUFYqFcNoQXFgsExX6Bi75qzahLK0AsV1XLsZeDRu4dMiz7EdTdpvJ6KT+HTMCvpma16mrb1n4niWPbA+FcOs6N3FgcEyChgqK5b+cB1zj5pUQ5t53MBgtEBxCZ52M/Bo80NkQV5j63cG7Ucn8emYFfSt0GrtbepD615l3rRDK0kZrvstmjOFrpIwFMY1l5Vy0tYzcTzau1IOP5vQ2mDPxPE1gqcIDDyatvBZkNfYtnvXZSKPncVo3Z10Cp92rKBPYqxWarVrtvSz7IH1HBgs8+SLOykJsbXol106v8qm76Ktf+8AJYGyAiFYRLTmn8YRPBoLmY0W5CEc2r3r0shjZ+F3J+1HRwr6tM09WqXVau2sooEnaGnXnDWTuUdNSrVQDQyWKQNPbdrJqs1vOm3w/XsHuO6845y0FXkSjuUFqAi7LshnZ1Gk3clYRUcK+rSM1aptmS2USylqySTRpgXBTY+8wFObdtZtgy/CJHQJ9CIvQGmQxyJVBLNBHjuLouxOxjI6UtDnzViNTlrbMZtXPP2CWT18/vwTWLX5zcw2+DgbfisRJdCLsADVi9G+SJnIY2dRlN3JWEZHCvo8GSuvSdss7SzqWePut2ZLP1cvf5rBYcW4LuHGS+a3LZkrSqCPZi1wNC9SLuTBu0XYnYxldKSgh/wYq5n17fPScLI+631r+xgYDjwGA8OKddt387WPndQQDfUiSqCPZi1wNC9SHp2JjhX0eaEZkzavXUK9i4UkfG4l4gR62gWsaE7b0bxIeXQmMgt6EbkQ+CZBmeJblVJft36/FvivjDQk+ZZS6tYG6Wwb6p20UQ7GFb27+NW2tyrJTvXuEhqppXPZ6dO5Z01fZfG67PTpme6dNxrZfRXVHu5NFcVFuxSDdiokmQS9iHQBNwMXEHSSWiUi9yulNliH3q2Uuj4nGtuOrJPWJXyASmkDs4xcV1d9u4RGauksmNXDnZ/uDI2z0+zhHtmRRYC2SzFot0KSVaM/E9iklOoFEJG7gEsBW9CPaUS1JzNj6SEwmVyxoL5686ZJCahk0qYVdkXXONNOXm8PH9vIKkDbpRi0WyHJKuiPAbYZn/uAsxzHXS4iHyToF/sFpdQ2xzEdiyjhY8fSj+8uMf/oydz82KbMmrU2Kd23to/vrd4GoXO13h1CK5Cl2mbaydtKe3izt95F8zWMBmQVoO1SDNqtkDTDGfsj4E6l1AER+VPgH4EP2weJyBJgCcDMmbXVGfNGKydRlPCxY+l1sTPdH/bKhTO4LENHKR1vXg61+UZ2CM1GFuGddfK2YnfS7K13u7f2oxVZBWi7HOXtdtBnFfSvADOMz9MZcboCYPWHvRX4W9eFlFLLgeUQNB7JSEcmtGMSuYSP/d3Nj20asbMPK+5YuZX71vZlos/VejAKzVzskq6dRXi3W/txodlb77TX91p/NeoRoO0yW7bTXJpV0K8CjheRYwkE/FXANeYBIjJNKfVq+PES4PmGqWwQ7baPRUELNO2grScKJy2jN3OxS3PtLMK73dqPC81efNJcv9O1/noXsaL7m4qATIJeKTUkItcDDxOEV96mlFovIsuA1Uqp+4HPicglwBDwJnBtzjRnRtIkqqcVXh7lDEw7+71r+hgerk+IpGH0Zi52aa6dVXgXbfI2e/FJc/2iKix5oNMXsSi0aoeW2UavlHoQeND6bqnx958Df944afkhbhJlZTCzA5SCmpLDWV+cFmiXnz69qS+8mRpp2msXTXhnRbPpT7p+o++wyGafTl7EotDKxW3MZMZGTaKsDBZXchio+8W1QojYi11eE78Z2q65a1q3fTc79xzg8EkTMjmriyzY6kEj43zHyq2Zci3SIM/xbYdfpt380crFbcwI+ii4GCyOAeJKDqd5ce1kLnMxsQub3bnk7IaFfd42fzu5DOCeNX3c+elkIdWppoB6xnnNln6W/nBdJddioAGhYi7AOmIsrwbzrfTLFIE/Wrm4jXlBD0FJAAn/h3it3GRIl40+yRfgunY9wr/RBcMubHbf2r7CCMIVvbucQh7Saz5j0RQQhRW9uxguj4xmSaQuoWLybylDw/u0aKVpryqzvE380crFbUwLelvwXhbayZOabkcxZNKLi8qYzapZ5KGNFKmwmQ2zH66NtJpPs7WlRhbaVu/qFs2ZwoRxwS60VBKWXTq/rvua/ItSlEqCoAoTApsFPRPHo9e+sqrtuZwFjbzPVi1uHSfoswy6S/AmNd3Oem/zO5fwiRL+Uc+wZks/Nz3yQsMmoqIVNjNh98M9efpkjjz0oEw2+mZqS40stHmbDNLwe15jYfNvlkJ69aJZi2L/3gGEIKS5FH6ul752m4DSoKMEfdZBjxK8WsiUpJYBohgvrpCZ+Z1rwpk09EwcH/kMroifLCYiEwtmJRc2a5c/oUag/P68usNXm0F3I2ahPE1KWUtFNGqT79870BLhbt67ESGa5GubMC45byGJ/0eLibCjBH3coLteWpSmE7Xlj2O8KM3c/u66846LtPknOXTNiJ8ScM5xU/n8+SekMhFFmZqimLKdmkorbZf1LGaNmIXyNCk1U8is2dJfye8YGAoK5wkwYVygyWv+bua7aeT5TKWoKzRXXXPWSKmVJB5Ly/9FzOJ2oWME/Zot/bzy1j66u0o1iUdxL80WdnEMEMd4cYXMssaXR53TM3E8JQk2nOO7S04hH0dL2nFc0buLJzbuYP9gMMHz0DyzCtNW2C7rXcwaWYjyXMSaJWSiop4UcCBjSexG0Mjzmb62obJi6Q/XMfeoSTVzvVFNvdXRQvWiIwS9OWG7S8JVZ86ssuVm1QyiGCCO8aJeeD11OFznrNnSz7IH1jNcVnSVhGvPnl2jVZkCtd5mKR+/dUVFwGuUSrVRGvVUojQLt0G0HyLN9fMwKzWiMTayEOW1iJm80jNxfG5atp0rYkKE3KNtotCIEF00ZwpdJamElJaVykRrlkWmldFC9aIjBL05YYfLiqMPO7hq4PPMKIxjPNcLr4cJXOeYk69cVtz685ertCqo9Qdcd95xkc/hoknfw8Z7px2a6I+IMpFVhbGFhdvuWb0NRBgadmvSSZp2Xmal0bLttuF6j3ma2cxxEQGFUA4VjE+9/1huf3pz6jGL82mlEeDmXMi6uJ/3O0fw6G92oMJ5kuX9NkNTHzUdpoqKuAmrB3fpxfNYt3136jDCuMQQW4DWiywv3nxGqG00ArX+gKzCUd/D1ugXn1FdRtrWhO9b2xc5VtrcpJQaKdw2HPwVVcQtSdPOyzY9WrbdJlzvMW9bvT0uUL37umDeUZl3c3beSL0hxVE296j7dpeEK8+YmSmj2hyHPH0eo6nDVCGhGfO+tX1Vgtx+4VqLTCoF3MzEkHozC81njGo0EqedphEGC2b1sPTieZUsypLAkg/MqZlQZnYwwD2rtzFcVjVjdd/aPr6/to/hsqqEshH+P64r0BJdtCZp2nlo4uYim9fC3Qq43mMzdiYu31XUb1loBVKFB7uulWRzd93XtcNvB9odndMRgl7j+2uDCAEtyKsG19IitRaalNiEUogEEQddDlu1jTgtvdEFRJtCohqNxGmnaYVB/94Bymrk+pMOHuekQy8Iw2XFcLjo2Ek0Qm37RAgihq5cOINjDjs40gSWZCKzNc4sXbrqKWRXFK3f9R6LujOxadWhw0nhwVHXSmtzL6JJrt00dYygT9J0ukKNfng4+PveNX2RNmL7PAWUh1XgiYqAGY7muq6d6FRWgc2zK2Nmoc0wZqOROE1LC+eH1r3KRfOnASPCUY/fojnupC4X9IJgCvHx46qTaCAotWCPv0l3lAMxSWvUv9ezJc6iXbV7y22jqELdRJRPK214sAsLZvWw7NL5VRE/cVFsjY5R3ot7u99bZkEvIhcC3ySoR3+rUurrEcddDtwLnKGUWt0QlQmICq2MsjVuf2sfdz6zNXKim+fpYxUwPOzeCbjC0czrRoWriQT276ztA+uNqNGmopW9uypmrO6SUAaGjeJmaa6/aM6USmKZxrVnz64x80TZeiEfB2I9W2LXYqYXal3zSF+j3VtuF+xFsBVtDtPy25ot/Vz9nRWVsb3z09U+LXPc0wp5jWvOmsncoyZlduJmRbPGs53ROZkEvYh0ATcDFxA0Bl8lIvcrpTZYx00C/j2wMi9Co+AKrZx39OQqTdEW4npSx2mtpsZoaqUujd0ORxOqt6Tm71W26jrth/UwTJQZSxc2g5HiZl/72EmxGq6eaPOOnsyv+nZXflv/6tuJtOq/zTaKjQjQerbELgXg6uVPV8bCrJDpMj/8xQ9+XbMg1Iu8w0QPDJYzFahLE8aaRehp8ykEZrvvG7SY415vOGgc72cJ+U0TfVakxb1RZNXozwQ2KaV6AUTkLuBSYIN13N8A/wX4Dw1TmADb8aIg0cmZpBXbjGBr9zYD2KYeu8l3lAkpTeq1fsY8k2tMGqDi1wXii5vZk/7as2dXCXptEspKTyM2y3p3OKbAuPmxTeHiF8B8t7ZwuvH+dc4FoR7kGSbaXRIGhgP+v3dNH5enWITS3D+t0NM8+8aeA1Xf2/4ZfW7e9X50kMJwQtntLNFnRbLxN4qsgv4YYJvxuQ84yzxARE4HZiil/q+IRAp6EVkCLAGYOdMdJpUGtgBb/8ruhtL/oxghbieQxXnYM3E8j23cwY6397P4jJk1x5oNInSkkH62uJCyJESZsXomjufGH61PVdzMnvSTDh7H1z52UsXun4W2PG2WjW6JF82ZwrguqQhwVzLcglk9sQtCPcgzTPTKhTO4Y+WIiTFtJEvS/dMIPXtX3d0lFYHralafx3Pb0WtmSLBddttUnNJGn6XlTbsWUFH9Jrk6Y0WkBPwdKfrEKqWWA8sBFi5cGFWVNhH6pWhH6HN9uysmkjRRMjbiGCGOAdI4D6HaRPD8a+urQsTsBhGDwwrt7kwKKUuDKDOKafeE6AiWqIiPRhafeid4nhNqwawe7lxyttNGbyJpQciKPDXHy06fnmiOrOf+aYSevatefObMmoiqpCquWeCKXrMhjmPHdwfBAvWUJXHxnVbKtCXBbitaJGQV9K8AM4zP08PvNCYB84HHJYhQOQq4X0QuaaZDVtvJh4bLVbHadpRMGiFh7xC2v7WPO1ZurVqtG9Hg4jTCFb12g4jgGYYdIWV5Cjxzx5K26Uo9900q41yvqSHrWEQVuEs6N+2CkBZ5j2eWa5mJhC5N1DYfxsEVCWZey9XGsJHntsOfSyUBpdA6/bguqSqzYSpu/XsHMt87qjKtqZRBEJgwMFjmpkdeyOxobjayCvpVwPEiciyBgL8KuEb/qJTaDUzVn0XkceBLzY66gRFmMyNbzC1slrR9c4egt8Pmag1uu3mSwHFphD0Tx1c06EVzahtEADUhZc2qab79rX2ptrX13CttGee4a7t2W1mv0ejYNWomyut6cSbGes91/W4mGtbj87J3qQND7iquWaDnup4nn3r/sbx9YKhioxfHsfYuNMu9o/jOVMo0ysBTm3ayavObiUmZrQy1zCTolVJDInI98DBBeOVtSqn1IrIMWK2Uur8ZRKaBKaDvWb2NoWFVlTUaZZJxbe10UwJzh2Bne9aT1m1rhPOOnlzjOF568TzuXrWVIw89qGKmsUPK8opYAZd9tbb6p3lsHlpYVNmGuCQ2iO4fkGUsOiWiopHnSDo3KkKrHp9X7S61vjaGGpoHrz17dqXe0+1Pb+by06dTDk0ow+WRnW8evqAoU9OEcdWKZbixSOW4bnVuRmYbvVLqQeBB67ulEceeWx9Z9UEP1r2rt2kCKr9FvSyTqQeMEqxa6A0NVTcB19metnBKow1rGvX3tsC+b20f967eFmr8u3n8hTcqUR16Mbn5sU30TBxfE/KXJTPURK191Z2x2ihzRo2/aSaLS2LTY5elf0AWOkYbGnmOpHMXzZlCd1f4XrqEUkKUmIZLEXDtUvNQSuzMckU0HzS6C4viu+9+ahE3PfICT23aGZqRAr+gUvFJkO1QNjomM1ZjRe8uhjKs7CbTi8E8ptAzPeoQNta26rwkacMuVE+oYBGJsuFH7Tyy1syxYda4d9lXzXHNwpz2pI+bLHGhqzZcDuUsGlseGl4R0MhzpDo3VJJKwI2/n9xVKs6UlNd4u2zzYvDt5WHP52a8V9disWBWD58//wRWbX6zssCk6cDVDmWj4wR93CBGvSwz9HHZAyOhhlFCz1XnJU4bjoXedSjFvKMn01WCodCrpBjpWetyKl133nGxZpw0yTBmjfulF0e37MvCnC7nmx77KJv/mi39lR62psktDdJobK6FZ7SjkeeIO9dWljSvxSEpWi3r+3HBVIy6u0vOBajV77WehawdykbHCfq0gxg18dOEGtp1XnQmbJokFfPe29/aVzOhFp8xk++u3Fq57rrtu53mGk2fSwAn1d3R0JNTEUT0xDVIthdEbWM3F5X71vaxc88B/vU3Oyp22YEsW1Nj0csTRatXU1SYMeFZNc5GtNSo3apz/ho80kiocZ6oZ8FttbLRcYIekgcxbuKbGmbUMT0TxyMiiFJ0dQmLrUzYOCQ5Pze+tqdyrGLENBQ1AeyFDYitu2OiZ+L4Sq2ashrZPcSNq76+HT1j5geYMJ1vcVpblMktD8Rpm0WqTNlMZC11kGSCSGuaS3P/KD+ZHeVmK0ZpeCSvsgijHaNe0KdhYPv3uAicW554iR1v7+fIQw+KPObG+9dVRRJoIZ+GWZKcnyt6d1Xq4QSRP6rSq3Pd9t187WMn1VzT5eCNqrtjQkcXKQJbbJxG76Jfa+tQ7VvQ6Dacb0ladTPtllHX7mRN346Dz1rqIM5cE2ePT7OIuN69NslIyd3TIE1UWNb7Zj0uLYq4aIxqQZ8lHtj8Pcrcsfjbv6jYx2E347oEsZpjrOitTnoaHlaZ4rkAN9ZLAAAgAElEQVSTkkt0lIKORCkzIuzT1DAxr++qu2OOzStv7WNcdzYHsmsXMPeoSVX5Ad2OXU6SM7eZdsuoazca/VDECQ21fH/56dNzKXWg0ZSQ1koPhCAsWvOkGeWW1Q+Wls48o2CKqjyMakGfKR7Y+H3BrOra7AtmBbHpQ1a71PdOO5SPzjuqJkrH5TCNihO3BUGSQLN///7avkw1TNIITFfFz7SmJ9cuYMGs5IzRYNyE8rCqajbeCidplEDO065clAkNtXwfF3qokWWhzTputvKx/a19rNnSX7XgDobKTLkMf3DG9Iowh5EoNxFh/tGTU5fcSEtnnrvJouZpjGpBn/SC7NBBU7jokMRVm99k7lGTAkEk1ZUcz55T22ZuwawePvw7R/KTDa9Xvlu3fTeXnz7d2U0nzg8QBdNPoKCidbsmSdz5UbDNR7pUctrIB73jsAu7xd1z42t7KjuhwWFV8UU0W1gm+WPq3UUUdUKDe9eYJvQw7UJbz7hddvp0du45EOSGPBM0iNe7zZ6J40cSE6FGmOsot7JSLHtgfWonbFo6XYpfvYiyFrR75zeqBX3ci4wLHXRN0uvOO46rzhyJeIGgvrotVNds6ad35ztVdIiDljzMAqbW/ZETj6xMkrietzr6Jak4l4sZ0wrdy06f7rx+HEM/tO7Vqs/Lf/YS5xw3NXdhadOQxmSU9hlM5KEF1isAks6Lmhd5OqDTLgpRSU4Dw4o7Vga8fPnp0ytNbHSkmQkd5RbHJ/YzmZ+TQkNdil8j4+IKjmj3zm9UC3qIT73WTkllhQ5GTVJdAXBgMMiGffLFnax8+c1KduodK7fyl//y6yqt3yygZNPSiCCwte59g8MMDccLxTVb+iMbaNhjZguCNGUV7MXgstOn15SLjWLoi+ZP48kXd1Y+b961l21vbqWrVOsHyYo4GrII5Cz1kBr1KdRr+kl7Xr2O0bzhSnJSYWiyLqugILaeftI7dEUMpUkiTKrxVO+CaI59nuVKGsGoF/RRiIu0WNHrrtqnJ++yH62vNNQYGAq65EBQXMwOLrly4YxMWlW99F80f1pVBp5LYNmO4iwZpmkEoq0dmxERSc3O9VZ8+c9eYvOuvUBoJisrZ1ewtIhLi9c7tbTvIS4aK0uUSRrUu+NrdKfYapOTzVdLL57Huu27uXdNX1X/YIFIX1TUXHIJ6gODZe5etTWT0uKK5skSsRPHW82MJsuCjhX0LuZIW3hs3jHVLfIUtcWZgEpEQxwNaSeRi2FcJpKH1r3KvGmHVglFU6NNUy89apt77dmzWf/q2zXX17CZ1oyIsFPSXffVwv4vfvDrynfDCnbsOcB9dZZxMAWXUgqRkbpEaf0HUc9nRlrlLRzrFQCNCo5WCR6Tx1xC2uUzMJ2udk6Hy8RmCupS6OhXBCbXpDDMqnc6rDj/vUdw6ozDYne4+rws8gSiTZ2tRMcJ+qgojjVb+rnpkRdSTdjLT5/Ovau3MWh1ydHV6gDOmN3Dly86MZcXZzOMWZnPNJEse2A9BwbLPPnizkrZZL1NPTAYOGt1yVbNWFCd3Ru1zdUJVgJV149zXgKVsgVRKek2zKgdjUd/swMVauJZe55qwaXNbbqwVFw5hyhEaY7NEI717vga3Sk2en4auASgK6jB9pVkcbraps35x0yuNB1SZcUVCWGYi+ZUt1984oU3+MyH3hP5zl3BFUkKgMvU2S5kFvQiciHwTYIyxbcqpb5u/f4Z4DpgGPgtsMRuHt4sRK2w+nvdbiwuiQhGygmbAk1rvFoAP/fKbue5cbRFTS6TYQYGyyx/srcSq34gbGQw890TqxKh9LEPrXu1IqSHyorvPNlbCZeEWkeQzZwPrXu16rrm9ZOcl2u29KdKSTefXUftmJm7qqwqPWK0jXb+0ZOrFo2o8dOCy6wiaPtkssCl/ccJx0Ycm/WafhoxGeVxfhLq3QGt37670q0p6TxbEC8+YyYbX0+uU6WxYFZ8+0X7nbueKUkBKFJkViZBLyJdwM3ABQT9YleJyP2WIL9DKXVLePwlBK0FL8yJ3lhEDeyK3l0VTTygi0SNT0+GJBtwIw40LST27BtECNufhdEHGoqgkcHK0I44aAp7YN60Q3n6pV2Vxg7DiqpohiTm1Lb/ikasxwicvg1ToK3oTS5b4Hp2s7GLDhudOeUQNu34LQBDQ9Vp8EnOtQWzRqoIRm39G4VLOBY5lj4OcREqedCfFDcfRdM9q7dVeDupsJ1r8bX7NsTBDl12CWr7ndsafpSvzzUO7S6JnVWjPxPYpJTqBRCRu4BLgYqgV0q9bRx/CLWN4JuGqLj5RXOCZB2z7ZfW+JKY3BU1EGeHdsG1AEFtTRqoreclUJUVuO3Nvfz8xZ2VrleTDh7HskvnV/WuNKMZbEaLmyB79g1y689frglJzZJhnObZdXehy8Mop3vX9PFSKORLQs2CeveqrZVxilpgs27980CRNLa0yBqhkiZc14bmMf1uk0KCYURpgIDnr1iQfC+X+Ser4lUSmH/MZBafMTNR8bOr3KaJfCpKSeysgv4YYJvxuQ84yz5IRK4DvgiMBz5cN3UZEBc3v2BWT0UYZm3J54oaSGNSiLuGuRVMWgXNRgbaV2BH32hhbWrJcYkyURPk5sc2VapymuaPKIGWhpHjFgO929KdvErAScdM5shDD+LxF96oaPvrX307laaXJt46TxRJYzOR1kyYFKGSNlzXBfPdpnkf9ljGBTnU++yuMSgreK5vN8+/uo7123fHLmbmPEm7wDfbTJYWTXHGKqVuBm4WkWuArwB/ZB8jIkuAJQAzZ6ZLaY7D99f2VbQ+l432mrNm1tWSL8k+m6ZZddQ1qpyIDoy3HJxAbGio1pJdgj0N9M7HLlOQJKzjbNZJ4/fKW/sqERJaqD/Xt5uuElx15kze2HOAn4ZZyEmaXqsFr6m5SsKxrcqOTFVAzHBCxkWoxIXr3rFyayWTNKokQdQO24Ws2q9rPNOa0jSfVOQF1QlcSSa4oi7wccgq6F8BZhifp4ffReEu4B9cPyillgPLARYuXNiQeSfKvucSOuYLzMKI9v3s+F3bJOOKuQaqwhZ12vW8aYdWImXmHT2Zddt312yVG02UiXoGc2dyyxMvVZUp+On611Jr7lE06meOWyR1vZ039hyolJYYKgdhl09s3FF5r0maXru2yjqXIEpItNKWn2ROsp2QcREqi+ZMqSlWt2jOFO5YubUSIquT4GxhH7fDtmFnsa7Z0u/sA5GUnJfWlGablrSZM40T2Dy/CCaZtMgq6FcBx4vIsQQC/irgGvMAETleKfVi+PH3gBdpMlz2PaBKkNhVHE1GLAmcfMzkSmJUnNZu2jRdiRZRzOa6zo0/CqIEnu7dFVvTXoeGag1koEGzRNQz7R+s3lssf7KXC+Yd5Vwko95DVEKVWeO8Z+L4SrSPWW/Hxo639zvfa1x/3DR0NqJdZy2x4BqXZpqUzHDTKKe0zgBPilBZMKuHPznnWL79s95K6C3UlrN4aN2rXHPWzKqxMU2TcVFQa7b0c/V3VlRoufH33T6DNEERWTRtcwdsmzzTKHz1mGTaWfMmk6BXSg2JyPXAwwThlbcppdaLyDJgtVLqfuB6ETkfGAT6cZht8obLvlcVsujYlpmMOKzgmc39PLO5v2KHBJxauy2gXGVT0yTdaNsoBGWIv2vQp4/X55qhoRDYFffsG0zVEDwqWsb1TDaUIpNQst+DmVClG0poh3GlAqaV3KTj8l0hc/OPnpy6REHceGTRrs3rQu1uLY1wiTsm78mvd4p/+cN1DJUVN/6o1imdZYd2689fruyohsLIKrucxUXzpzmVhzRCVysCEPDKbU+97HS8u4IiUKpqMatH044zeWZB0ntsd4RWZhu9UupB4EHru6XG3/8+B7oyIc4GbtrhXKu/HfUyMFRm2Y/W8/xrexgartXaL5o/jZWh7bKrNJJMpbXXy06fnirp5shDDwKqY/FdWrAOkTQhUJVQlbWhQlSI5WB4H22WHdclqbSbqPegx2VwaKTxuhmvXwLOOW4qnz//hAr9d346OmQu7W4pbjzSJs25xs8VrpqmxEIUf6ahu56FQMejw0gJD5fvKc0OzcwG193C9Hmmjd72d/XvHUjlv7Bttr1v/NbpeLd51sxpMSOs2qFpp3mP7Y7Q6pjMWPsF68l1yxMvVTIv7bT4pRfP47af9/LSG+9UMZxZ/mBoWHHVWTOq6mMT2vURYeNre7jx/nU1kQmuTEBbCD6+cUflPK3ZmlqwGSKpnbYSHpsmuiQpWkZPwrlHTaoKHdMmpUoWUwPvoSYkLXwOnXlrCnnXhLOvl7ZEgf7e9EGY4ax6vHsmjo+0B9uLgitc1UWjKyTRJYCSJn+9WqAtPOt1gC2aEyS3DQyWKRndwiCwyZt2+ahdi7aBR0XsmFnoJRmh1Xa82/NnRe+umk5U9QjrPDTtOAVE09RuB27HCHrAGQnw5ItvUHY4hNZs6a8S0FGw62P/xx/8ujLph4cDs0e9hcR09m3PxPFVkTW2/VTbEe9eFfSPRaC7lNwZKsleazsRrzvvuKABy/DI1tkuR5BV+zGfWWvn5vPq3+5YubUmSSqpP27UbikuXd3cSVw0f1qsPdheFNLUdc8Skpg0+evVAqNKeGRFFlOI69j/+INfV5llonYW5jxY9kB1dmsUzLHrKgn3rumr7MCjuqq5kIem7XqPUYmCo8JGX1ToiBEdhqfth/17ByIdQnboWBRKUp1cZUf3aFOOuWD8attbkZmA9mJkrvxQHRGwc8+BijlIoLKNHi7DR37ncE4xijC5oHct2ja+9IfrgEAbczlO9WSLKhmbpP0kLQJR2+o1W/pZGtqUIYjvthtEZxEyrslrT8bPn39C5CQ3FwUBZr57Iks++J6qBcem37x3loU/bvLXqwWawrNRoRL3zpJ2X66dRdJ5UdmtcYJz+1v7uPOZrZH+uDjkoWnbO2SITxRsB0a9oNdee9uO/dC6V/n8+SdUabSmA9MOHXNBmxdME4EdBaI1/a/8y6+DOivATza8HjQJsTQ5V1ja3KMmOUMS7129rUorPPeEw6tomzppQmwomkb/3oGKA3QoFPa6o5ZLIxrfXeLcuUfw0w2vV3Yt963t4761fax/ZXdsck3SIqAnwjyrjs331/ZVZS1LhGkq7h5JJp44P449yc2dUBnY+ubeyExblxMyTQVRjTibchaNOst160VSeKON+UdPrvp86ITuVNVj05ohteDUvBXlj4tDI2Nsw9whp3VGtwqjXtCbXnsT86YdWqPR3vKz3qqqjGaf03lHT+buVVsr9nnTUQhBSN+efYOVOjSKEUbu3ztQU7rAxWiusDS967Dty7ZWOHXSBMZ3SdV2PG1mr1n+oaxUjRPR1Ii0WaqrFDhPu0rC91ZvY8igx4yUicspMBcB05yhr6EF4z2rR5Ktu0rCp99/LLc/vTmVLT6Lphzlx3EdZxdKi7K92jT17x1w9s81hWRShU/7WZqlBWYxw9VT86l/70Clc1RJguSsLGaStDZuU6M26yfZNXbiCuNBdY5LVrj4oEix9qNe0Lv0cQFuf3ozF8w7qpIWr4+L20pp7drc3gNVDG43xobaxBJwa3KusLS5R01yMrCtFbrsw2kze13lH/RvWgjdE9p0RYKSrTq/YOaUQyp1aDTMBTCueYOGy0xmhnaau6TFZ8zght89kQvmHZVoi48LU0wrIM0xMHdGC2aNFEqLs726NDf73i6bf1bTV95hmFmdkK7wxqSaT3HRXfV0+4oTnHrMzbh4s8YORLf0y8Mh6+LNZi7SWTHqBf3lp0+vaJyl0EBmCj6d/arbl8VFWyTZ2gg1XFcEj9bidu45wOGTJjidQXOPmsQF7z2SHW/vZ/EZI1ELLgZ2aYX6Xhpp7Yuu8g81CCOJFFJxxg4rKhUlNcxIGXOhicop0HTaC6F+D/bk1zuVKM0rbZiifndZtVU7m9m+X1QYYdy9TJs/uEtAuxYQnSE97+jJqUwlWZBmdxSnUcdVbdRwjV/aCpMmfTpBMI2NW/tY7Bo7QOTz5uGQzdME1AyMekEPgXY9UuZXKIWahrYlau10wawejj9yUuzE2fjaHr63ahtlpZy2tigGT1q97YkM1GiQJtJqAx84/vDKwhF3vHk9V3anFu6qXN3XU4/vSdMnM/+YyVUt/+zJH5dhaZvJzDE0Jz/EN1N2jYvLsWzmImTRVu3GJ/a4mbV5zHK1cZPbtvnbSWI2DQOD5arexFq5qEcQRS2aSUpCXlEjLnNZ1LkmrT0Tx1fMpGVFprLTUc8W9bx5hT4WSYO3MeoFvXaQag0UpegOQynNqJthBas291cahkTFvZrRHwM52trsiZwlqsQF2+79/GvpyvK6JrBLW7P7ei79/XlArRDOEn5nCkwtjO3fsjZTdglfOxch6Ro9E8dXdnB2pJFr3HRtnrSatqntRdnozXcgIlWJSsPlgKezlsdOcl7HvbtWR43Y43vitEOdZtI0iHq2qOctujaeB0a9oNcTxMxwLZdVJYSxpkrdYHTiy4re2r6wPRPH5+KssSeydmYNhB2kzMQhE1EaWZYwPhNRE9jF6Gl8Alknf1KceZZCcy7he5mRqZzWFrzsgfVVTd/tbkP2uOnaPC5HepwZI2khrEpaM3I8xnena9NoI8kkEUeTHZWVtnlIvcKyShEaViNBEVbkW9r7Zd0lF1kbzwOjXtCbdvV71/QxNBRsj5/atJNVm99k6cXzeGzjjkqMvU6A0k4bM/dz0ZzqNndKwbIH1gM0bCO1J7KZJappdTnntBNP94OddPA4Z3ho2kbgUdvUpIlhas5DQ/V3cYpboLJUPNTXsoVvkvYWdQ2NqDaTWc0B9cAcb91foJGm0o2YJOx5ldQ8pFGHpkthc+Uw5HW/PNHOYmVpMeoFPVR73O2QuP69A5w64zD+9fnXK2FeehvoKi/rCquzOxzVm25tT+So8D2NFb27KsXMhhLCQ13CIGoyZN2mVofWgYShl67Y8iSmj1ugTKdlUsVDvWDaGbFxfg8XbM01KqsyqzmgUeShYTZqkjD9N/WW28hKq62wReUwZL1fs4SxrYwtu3R+ZH3+dqIjBL3GgllBSNzKl4MoDrMokq3ZRDGKvoaOBLE7HImVXFSvJmHeJ6pEgUtjzmI2iXvGLDSb11EqiM2B+pKmVvTu4sZL5jtr7qcx27iiU3Tp43p2XVmEYRpzQNG0u0YWDJf/I20oZVQNoTS0uhQ2W5Bn2a00U/tf0burogSaCYlFePcmOkrQV6Czl8L/TW1Bm2rSJGCs6B1JJoJgKzlv2qH8+pXduVWh+8Dxh/Pob3Y4NeT+vQMVh5SGK2IjCvU2VrFhR41o2G394qJX0iwCacw2rsQUXaOnXo0yL/tskcwJjSLK/xFn464xTdY5DrayFdW4O6t5rtE561rEF81xJyQW7b2Xsp4gIheKyEYR2SQiNzh+/6KIbBCR50TkX0VkVj6kpkNVFE5YP1vj+2sDW+PHb10BBNvuL350bmQq9nXnHcdlp09nfHeJLoEJ44Ia6fpz1giImx/bxJot/ZXPH791BY9seJ3hcnXonIY2c2h0leDqM2emmji6aNtQWaEUXHv27MhzbNps6El1zvFTK7kKrrZ+i+YEbepgJHpFX9M12UykNdvoxcte8PRilPW95ImkZ0yLuPeR9K4aubYJramWVVDB1fR/JGH99t2VcwcG6xsHzXNR8zPLzikP3tDz9b/9ZCMfv3VFZfwWzAoSErtLEuk41mN+x8qtDb27RpBJoxeRLuBm4AKCxuCrROR+pdQG47BfAguVUntF5LPA3wKL8yI4CVGaer3hYgtmjbT804XITEdZGrg0PTuJxuUEXDCruu0bitQT7r61fRU7uAK+8/OXmTnlkJrIjbRaqG1qKpWkppaJTa8ZvZK01U6zFY/T+hu1R+eBPOKx495HozuGLOf3TBw/EsVGchy77ccxz03bJMdG1E4r6zjkwRtRi7i+5t1/6i4ilzYrutnIaro5E9iklOoFEJG7gEuBiqBXSj1mHL8C+ESjRGZB1EvNEi5mF+DS29BVm99k7lGTgFpHLkRnYrqYJK0T0G775tIWzGQj/be9COnqleUwz0DfL+u2Ns7UZNLrKiSXlMIelfWqv7O1/nXbd9c4X9u5ZW6WQNHXifotbemELO/arlNj7rDs+63ZUt3q0qz7FNUkJ01z8XrGyEWjPqcRBcDlg7AXG7sHhUlrXFZ0K5BV0B8DbDM+9wFnxRz/SeChrEQ1iiinWZpwMTvO285KvG9tH9ve3JspE9Ol6cUJBXsixR1n2lERqTiJl148j+4uqRQj00XK7FKuaavs2ZoJuCOQ9A4oqpCcazJEvTf9LnQhtxsvmR9ZcbMo9vBGF5u4XYHrtyw7gCwVFaPoMPlAZ5s/27e7KkRVEdRqKpdVVc6IFnAbX9uT2Fy8njFyVde050UjwRNJpbBd17X9W1l8bHmiac5YEfkEsBD4UMTvS4AlADNnNh6OFKXVRKX7x4U0mnHeZlaiCNy9ahvl8kjf066SsHPPAWefS40oYe0SCnF1V2xUMVuQFlyhoX/vAHdbZQeWPbC+ppTruu27K/Xu45xtZnQBxrO7hG1cITlTA9VZonv2DbL+1bdrtDvT/DQwHGjwURU3i+gAqwdxC7v9G1DTBctceF2O6yw7DhdfmHygey3bKAlcuTCoe2TmjOgdnlmtFEaai2cZI9OcqnnKWV3TmheN8Ik9D9MsmraTOmvSW17IKuhfAWYYn6eH31UhbA7+H4EPKaUOuC6klFoOLAdYuHChch2ThKT62HbXIle6v8tWbMZ566zEddt3j3R40s9AYIN8dOOOmj6XWaop1ru9ts0/iFSFwdn31L4Fs5SrKah1HRsXI5o2W4AL3nskQKVuvUmrS4vR5rI7Vm6tWnBM2NqdbX4SqqtNps1+LTqy8Ir5/GbTeNfCm6ayZhQ9prKhs42hlg9s6N2bXULC3OHZuGj+tFh6XPTZ5lRz3pjVNV3zIg9kMdO126QI2QX9KuB4ETmWQMBfBVxjHiAipwHfBi5USu3IhUoH7BVcN9fQle6Amro12gGbZCt2JSLd/NgmyuVaFh82tH8diQLxhbminsM1OePikV0anl74zGbl9g5Clzao6swTUX9HC6Htb+2rstlOnTSBe61uW5oGW4vRdXPufGZrRduKEhamdnfZ6dO5Z82IMDcFTqudr81OuMnqYNWatYYInDv3CB4JEwPr0eDNdx2lbLiiobq7hMULZ9QUq9Owd3gQzJVZU4Ks1zxs9LYCZxYf1Ock7fazIu2iWYS8ikyCXik1JCLXAw8DXcBtSqn1IrIMWK2Uuh/4r8C7gHskaC69VSl1Sc50W0k81WYCXVHQ1cEe0tUdsX9fNGekUbIIlEpS6UVragyXZ3RwmscODAb12XVRsZ17DnDjj9bH2hddtCb1LHVpxaYtVcfAQ3S9eYGabluuewCVPrRlRaWZSUkpatvFBHkK5sJ256cXRUY4Jb3HvCZZsxNu6onxtuO3IVh87YADINYvomH7e6KSpBbNmcJBxjz4yIlH8qcfek/lWVxj7drhje8u8d/+4NS6xjGrzwtq61O5dvt581KR8ioy2+iVUg8CD1rfLTX+Pj8HuhJhvmwxNHpd6c4UzHYHexNpX6BLS9UaP9QyeVanl1n3ZmXvLhBhcKja8ZlGCNg+hrjz7GfShbR0DLzpexguKz584hHsHxyuNEyxm5jHlcTVfWgh0D6vOmtmRQN88fU9PLvtLU6dcRi3P725pq68q1RFEu5YuZW/DNs7jusS7lxydt2TLKswrifGO6sJasGs2oYyZiP5NPVpop5xaFhx0vRDmX/M5Bq/jUug2ouEHUGWt506jc8rKRLJtdtPu+vOsuvKI1ErD4zazFibeczu8WlWeIhOp48K79Pf2fZLW7PMar+rqa9jOJAgutCWC7aPIUvP0vXbd1di4IeGyvzr869XlX544oU3GBouVwqwmeNfqVHi2H0smFUdX6/CAmTXnDWTNVv6+R+PvsjAUJm+/n01ERqQXHLYFe5n1nMfGFZVWbpxiMp+bFa6fb0mqDVb+unfO8CyS+fX8GxSwIHrWmbBujLw61d2s/H1PVXmMpNm83pVu9KI5tx526mTfF5x7yBut+9CI7uuPIveNYJRK+ihutejS0gnMZdtNrG3clBra0/70k3zSFKyyIJZtfV1tDkoLsY+6lpJxc5MmILNjNnXphyILv2gTQKusEt7XKLyAaKcaGkrREYlo9k939Mkt8VFPKUVxt83mlTnqf3bceFxgqzeham7JJw0fXLNe4b4OHR9v3qaczcDSXM07W7fPL7eXVe7k/g0RrWgd1WOS+sc6Zk4vqpgkyveF2q1yayT6OrvjPSgvfPT0YlVmil0O0KNqLaEcUirPbkEW9QuafEZM9n4+vpIQZ20+4hi+jgnmj4ma4OMRXOmMN7c1XSJUzO1x8IOVTQFRFrH2z0RDmqXQE+r/dvHXX769MQ68/XUgRkuK+YdM5mNr++JTQxy2b+/+6lF3PLESzz6mx01rTaTkLfDMmmOZhXAjQjsvHcy9WJUC3ozpjdN5ThXOrLdLchmjkYSnbRtGYIF49tPvMTPXnwjdtLcu3pbTW/VLLbprONnFyH72sdOqtzH7u/p6vdph3jG7T5cTL9gVhATffeqrRx56EHO9xc3WaIcc2l2NXZ4rskX9Wy1V/TucjqoowR62t2hfZwi2QeUVsDY42c3oc9itnjyxTcqAQpJvQQ0muGwbIcmXZTomiiMakFvRx4kVY4zF4bg+JGGFVENtLMkOtnaux1C+Prb+2Mnje1I1TQ2axu8Z99gpS+nAu5Zva0q/tk0jenPUdpcVvODqdUuvX9dmMG7m8dfeKMmSigOUfdPEnSuBBvtzD/nuKmRHb/i4BKaEG1KSLs7TBLGjfBF3PhppNFP1yIAABwnSURBVKHR3NnFFaWLOi9vPm/Ehp/1+DS+vnZjVAv6BbNqIw/itDA72cM2M0QJsjQvy9bev7826Dt67+ptlRT+KPOHhu1IhcZSppMiD279+ctV3w0NVy+UWQqepTFruK717SdeqpRpACrRNc3eHsf5BuoR8pqONOYpV75Bkg8nSRg3grjxS0tjvXbsrKbQPBa3rItL3PG2yS+PftDNwKgW9ECkJq5hModZqEmA9x9fn+bmgq29K0Yco0nmDw3b5BCVgJIGWSMPoHZByVPbcl0L4JGwxaOJu1dtY/7Rk+vq1JNWGKTxDdSDrLuetIuUfVyzTQWubN0k+upJzlo0J7rYXRYHdBZkXZSijneZgksOX58X9DkhihGTsk4/f/4JAHWVULVha+96227TljRpTHNJI51qskQe6MSXc+ceUWWmySM8TCdl7dxzoCYJ5761fc6kqeEU/paoe6UVBlow6UW1nrHOImzr2XXE3beZiTj1Xj/tM7qubyd1ZXVAZ0HWRSnqeNNcpU1+F82f5vT1tRsdIeijYAs7OyUcqKqO2EhSjUt7rwd5TeKskQfg1pgacWqt2dLP4m//gtCiRXdXdaeibz/xUtXxZjetejr1ZA1t1OdkTcbSz9aurMdm2bWLdP16HNBZkHXhdR2v55gO0zT7VRTNMdvRgj4qIkMP/l/84NdV1RGz2oZt5KG15TXJ0ghpk96oVnxp7e+u+9y3tq8i5CGoC2Q2Tpk6aULVdRbO7uGXW99K5W9x0WCGNkopPgkGasc6S9P3ZgvDOOSx0yr69ZvpgK4HLh5fMGukJLfdn6EoAl6jowV9krBzVUeE9CWPm2EnzXOSZWG4eu8bp9ma+QAQlD4wr2ubu2646ESgviYRdsSSsp0mDtihoVnq2zdbGMah0Z3WaLm+q0xyKwVoUnVcGCnYVjSbvI2OFvQQL+xc1RGjQqXsl7304nnc+KMRW1yWkMAkeuudZHbXnqwLUZq69DbiNNvDLY39/BOPrPFXuMxd9ZqqzFBblcL0Y4511vr2ce+pFTHVzdYazesndWvSv0f1FUhDv/bl2F3d7DLJrUJkfXuLN0zzjYgktlxsFzpW0KeZbAtmBdURzeNME8b+wTJ/+cN1KKWqSyEPlrl71daacMq8Jl49k/iOlVuruvZs3fVOTYGwqGvG1R9PQpxmay+kusqh61nTlIqIw4JZ2UJtXffPWt/e9Z5aabtvxYIS18XMLBOy3yiZnLVrlO4kFtXVLQ8tOetYxYXfmrwRZ74pEjILehG5EPgmQZniW5VSX7d+/yBwE3AycJVS6t48CM2CrNEX5m+L5kwJShCHTDdsaIgVRyEwobtUdZ0UloKm4qF1r1Z9/vH612qyXqPGIIu92RV2Fxc6aC6k4I5wyks4NuIIcz1HPYK0Vbb7Vi0oVc9jFNsbGCxz0yMvMPPdEysKj4ksXaNss5vZ1S0Ps1g9Y5Ul/HY0mG8yCXoR6QJuBi4g6Be7SkTuV0ptMA7bClwLfCkvIrOi0cmmrPhy3b2nUgpZ4LgjJ/HstrdqwinbhYvmT6toUgCnzjgsaCBhlB22u/5opLU3xxX9SlpI4yZbnsKxEZOGba6oR5C2ynbfqgWl2jQR9GEYHlYjJbXD2vW2sM/SNcpOFBzfHXR1yyu7tJ6xymJCbae/Ji2yavRnApuUUr0AInIXcClQEfRKqc3hb64Q6ZagZ+J4ShIE62Ud+BW9uzDlfJfgrIXT7igAG9ecNZOtu95h+ZO9KBVo9OfOPaLS6m94uJbB0yStmDAnjNbo0iacxU22NBOl1bVE6hWkzXZkauQpXOLG1jZNCFRVuNR9Ch77zQ6GyoqSwJIPzMncAzZLxdWsz1DvWKVVGlr1zhtBVkF/DGB29u0DzsqPnMah+0kOZyyupGEmEukSppppXWaBIr3USQePA0bKxE6dNIEJ49wMniZpxYap3WmNTtemz7oVtu2ccROlHTHrjQjSZjtK9T3MhK96kWZsTdOEq8LlEZMmVJW01nyY9XnqHbM0z1BPoEEWtOKdN4K2OWNFZAmwBGDmzOyp7lEws9WyFFfSSLI5F/llZok9bmQ7W9UkJeJcM4pCT66kVm95+BDyQt5aWrN2JPUmfGmkGdskvgKa1qg9zbiZz2D7o6ICDYpebTJvZBX0rwAzjM/Tw+8yQym1HFgOsHDhwtx8mS5NLOtLLbpAj0KUcGrELu+6h9kkJaoZSFTf2nrGtV020Lz4oFk7kjwWwDRjm4avGqlgGndsmnFbNGekVaXtj3KNEeRXN2e0IKugXwUcLyLHEgj4q4BrcqeqAdhMCWPrpeZhV7Tj8bOcC9n61qZ9pqLbQOPQrB1JHguga764IqOS+CoN32Vd8NKO24JZ1a0qTX+Ua4zamdXcLmQS9EqpIRG5HniYILzyNqXUehFZBqxWSt0vImcAPwB6gN8Xkb9WSs3LnfIYmEwXldrv4Z6cdjw+uOOh4ya2HUXRjtokRUKzdiR5LYBpIqPyQFYBm2XcolpVRo1R0aNk8kZmG71S6kHgQeu7pcbfqwhMOoVAltDBdmqM7b6/hh2PnyUeWiOPKIpOQjN3JHkugM3WdLMueFnGLYtvbbTvEOuBqDRFQZqMhQsXqtWrVzft+klCtJ2VCItwfxOmRg/wtY+dVFdd+KKhKAtpkaH5UAviZvChfw/5QkTWKKUWJh3XsSUQTCRpPe222bX7/ia0UI+z0Y82FGkhLTJaoekW0QQ3FhafMSHok9DuzLZm3L8R5r3mrJkdIeA1irSQFh1FFMTNRKuUgKQAh2bDC3rab7NrRsy212BH0O6F3KO4aIUSkDbAoZnwgj5EuzWZ0eRUG21o90LuUVy0QgnII8ChUXhB34HwGmwt2r2QexQTLiUgb5u9XXAwS8G3vDAmom7GIsaCg8nDI280y+zZaFOgKPiomzEOr8F6eGRHs8yeZoBDO3xopeRDPDw8PMYGtNmzS/LJ6HYhqv5OM+E1eg8PD48QrXDct8OH5m30Hh4eHi2Gt9F7eHh4dDha7UPzNnoPDw+PDocX9B4eHh4djsyCXkQuFJGNIrJJRG5w/D5BRO4Of18pIrPzINTDw8PDoz5kEvQi0gXcDFwEvBe4WkTeax32SaBfKXUc8N+B/5IHoR4eHh4e9SGrM/ZMYJNSqhdARO4CLgU2GMdcCtwY/n0v8C0REdXE8J41W/q55YmXePmN3zLn8Hfxpx96TyWV+ZYnXuKJjTsYGFZMHFfi3549m5d2vlM5dv/gMM+9sptzTzicM4+dwm0/72XfUJl50w7l3LlH8NjGHex4ez9nz5nC2weGEGDe0ZNZt313VVMN3Qx7554DFboOnzShpumGq2k2wNcffJ4fr3+NC+cdxQ2/e2LNseY95x09mf69A/RMHF9Fx8bX9nD3qq0ceehBlTHQcFXP09ff9PoeDgyVWXzGzJqsvZ+uf40fr3+Nme+eyO59g0zoLnH8kZMqTZbNFnR2GrkeD3scTFq27nqH//3UywyWFYcdPI53HzKeOYe/i3PnHuF8RtuBlSZ6od4IhzR9DPTvG1/bE5v56LrWHSu3Rr4vfXzPxPH07x1gz75B1r/6dk31wyga7fNd78h8F1MOGc/LO9+per/2M5k0T+gucdjE8U4ejxojoIr3XfS46E4qS7BmSz9ff+h5tr25lxOOnISCxCqR5n3i+Esjqvqk/XxZ+KyV2euZwitF5ArgQqXUp8LPfwicpZS63jhmXXhMX/j5pfCYna5rQmPhlWu29LP4279gqDzy3bgu4a8vmc/SH/666vtmYHx3iRt/fx433r+u0jrP/l03xrabZuvffrr+NW75WW/lnM98cA4XzDuq6lgbApi/dJVg2BqDu5aczYJZPc5mInOPmuS8/mc+OIfbn97MwFAZASJuT3eXUBJhaLhMd0kg/Ht8d4mlF9eOh37Wja/tqaIlDvYzmmMJ6TIM681CTDrP/L0kVPGZOYZ6PJY9sL7qWvY4mO9LX/vAYBnX8OtmMFE02ueXhJp35KLBhM1PX/vYSQDO4+334hqj7pJQBoZCnjD5R9MD1NAdNX4mD/zBLb9w8mlU05yo8Y16jqhmPPbz2eMbx2d5ZcemDa9smzNWRJaIyGoRWf3GG2/UfZ0VvbtqhPngsOKhda82XchDkNn20LpXq5ph27/rzLeoptk/Xv9a1Tk/Xv9azbE27F+GHWOg7+uqnhd1/R+vf62StRdze4aGFYM6u8/8O2I89LPatMTBvr2dRZgmw7DeLMSk88zfbT4zx1CPh30texzM96WvHTX8+twoGu3zXe8o6V3Y/PTQulcjj0819sOqIuTB4p/wfCfdEeNn3iOKT6PojRrfqOdwzR/X89nPE4dWZ8dmFfSvADOMz9PD75zHiEg3MBmoeQql1HKl1EKl1MLDDz88IxkjWDRnCt3WU4zrEi6aP63m+2ZgXHeJi+ZPY1yXRP6ut3W6abb924Xzjqo658J5R9Uca8P+pcsxBvq+drW8i+ZPi7z+hfOOqqSAx9ye7i5hnE4VN/+OGA/9rFkq99m3t7MI06Sr15vSnnSe+bvNZ+YY6vGwr2WPg/m+9LWj2FefG0WjfX7J8Y6S3oXNTxfNnxZ5fKqx7xK6DZ7odtDjpDti/Mx7RPFpFL1R4xv1HK7543o++3ni0IpSCyaymm66gReAjxAI9FXANUqp9cYx1wEnKaU+IyJXAZcppf4g7rqNZsZ6G7230XsbvbfRj0UbfVrTTeYSCCLyu8BNQBdwm1LqP4nIMmC1Uup+ETkI+GfgNOBN4CrtvI2CL4Hg4eHhkR1NK4GglHoQeND6bqnx937gyqzX9fDw8PBoDnxmrIeHh0eHwwt6Dw8Pjw6HF/QeHh4eHY5C1KMXkTeALXWePhWITMZqA4pEj6clGkWix9MSDU9PNKYChyilEuPTCyHoG4GIrE7jdW4VikSPpyUaRaLH0xINT080stDiTTceHh4eHQ4v6D08PDw6HJ0g6Je3mwALRaLH0xKNItHjaYmGpycaqWkZ9TZ6Dw8PD494dIJG7+Hh4eERAy/oM0JEYmo6jm34sYmGH5to+LGJRl5j4wV9dhwGlUqebYeIXCMip4R/t3vCHKT/KAAtRUNh+KZgPAMF5BsRKYpsfBdU2rjWjaI8jBMi8m9E5G/aTQeAiEwWkYeBHwMopYbaTM/5IvIkQSXR00Ka2uJwEZGPisgvCNpGfrydtIT0eL5x01IYngnpKRrfXCIiX2zX/Q06RESOEJHHgVsBlFLDjVyzcII+fMguEfkU8A3gBhH5QLvpAvYBbwHzReRKaHyVzYpwbA4Wke8BXwG+StCXd2I76AnveTiwDPhb4LvAYhH58/C3lvGX5xs3isgz4X0LwTfh/bpF5MvA3wPfEJFTlVLldo1NuNjtD/+dLCIXhXTWPy5KqUL+A84FJgGfBh5vMy1dwJHAF4CLgdeM36QN9Fxq/P0J4Ok2jYsA84FvG9+9l6APwdR2jI/nm2LzTIH55t8QmJA+D6xsM9+UwvH4OnBpHu+qMBq9iHxORL4TamQATyil9iilvgMcIiKfDI9rOs0GLX8iIqKCbdPbwO8ppR4AnhORpSIyXymlmm1XNOj5NIBS6ofh913Ay8B6EZkRd40cafkjEbkgpEMBvwXeJyLvDr/bAHwP+B8tosfzTTwtbeeZ8L5F5Juvi4jufvd/lVL7lVI3AUeIyDXhceNaSMvlAEqpMrAdOAF4CnhVRD4jIsfXfZN2rlzGCnYtsAK4EHgC+HPgPcbvFwHrgZ420PIXwHuAI4Cvhsf8CTBE0FULYFyL6Zlj/H4SQUvHSU0elx6CLf+rwHNAl/HbPwH/bB27EjjW803r+aYoPFNEviHYTXyBQIBeATwfjtcRxjEfA15pwdhE0fJuYCHwV+FxXwLeAX4Ufu7Oeq+iaPQfAf6LUurHwP9LsIX6uP5RKfUQwSAsEZFJ2tbZIlomEHTM2gdcJCI/AT4HPMpIxc1mOthsesYTbL0BUEr9msCWd1UTaUAp1Q/8BDgRWAMsNX6+HrhQRM4IP78D/AoYaCZNeL5JS0tbeCa8V6H4RgWS8jzgK0qpewkE7cnA/2Mc8wPgBRH5EgRO7BbScipwAfAa8AEReRD4Y4LFQLdkzeyYbaugN7bTvySwYaKUWg08DRwjIucYh38Z+M/Ai8BRLaZlDvB+4KfAM0qpU5VSHwXOFZFjwxfWKnpWEIzN+8PjBHgYOKhZpgDjuv+klHoL+J/AZSIyK6TrbeCvgb8UkT8icPrNI9ieN4MezzfZaGk5zxj3gTbxjf1sxvisBj4Q0vBjAt6YJyJzjcM/C/ytiLwGHNNCWjYCpxBERfUBq5RS8wgW5XNF5Jh6+KbV3u15EjQPByq2KAhWq5KIfDD8vI5gq3d0eN5xBEzyL8DpSqmG7XgZaFlPMOCTgKVKqa8Yl5mplHq5UVoy0rOOwH43LTxOEZgH3slLcDhoUeH/+8P/VwEPAf/JOOZbBGF7C4BZwJVKqd050XOOiLzHuFc7+SYtLU3nm4zj0lSeiaCnrXwDHGx+MMZnEzBJRE4KPz8BTCZ4V4jIqcB3gPsI+OYfW0jLz0I6dgCfUUr9VXj8m8A5SqlX6rl5SwS9iJwsIj8nCO2aYnyv7/8iwcRYLCJdSqk+gmiF2eHvu4HrlVKXKaW2t5iWbQSCY5ZSakCCEL4SgFLqnUZoqZOePgLNdLZxmS8ppW5rIi3icGZ+CzguXBSOFJHjlFKPAl9QSv1Ro+8pvO/pocnjUYKJqL9vB99kpaVpfFPnuDSFZxLoaRffLBKR+4CbJYjV7wq/18lqzxCYzT4qIt0qcAQfQ2AXB9gF/Dul1JU58E1WWtYTLHinKaX2h3wjAEqpunc6rdLovwLcq5T6mF6RQgbUq9oe4EkCu+Y3JPB09xAMOEqpN5RSL7aRlsMMWoaNY9tFT2VsQprysmlG0aJUEFd8sIi8K7znVuAHwK8JNKJDw+8bSuwI7zlORL5NUJ3v7wnMDOca9LSMbxqkJVe+yWtcQnoa5pkU9LSUb8L7nkuwi/s+gRnkE0CPiJRUmKymlNpEYDJ5D3BDeOoBQt+JUmqbCvwY7aRlc/j7cC67LtVcr3KJwE55m/HdBQQToDv8/FXgHuB3CLaWtxPYFL+N4aHvJFqKRk9KWv6GgGFPDj9fTTAx/paco0cI0r4/Dhwcfr6WYCvdbRzz1y0aG09LY/T8Vav4Jrz+5wjj84HpwJ3AQcbvXwX+F8Hu5neA+wmcxN8GSh1LSxMGehFwgvH5UIKt5MUEttKHCcKq/jx8wDuA44zjS+QU9lUkWopGTw60LCLHMDiTHqxkGeCTwC36N4IoiTuoDqVsyth4WnKnp2l8E34+lSDx6q+A14HHgduAxcD7HHz8LuCwTqOlhrYcB/ww4P8SbBu/QtC0Vv/2F8Ba4JLw8weBHwJnmwzZibQUjZ4caMlbK3TSEwqKUvj3ceFE6dG/tXJsxjotOdHTbL55l/HbmQQC9fLw8ycJdhqntOhdtY2WqH952ugPIdAC/yz8+4PGbw8QaIXvDj+vJogT3Q+BE0nla/cuEi1Fo6dRWnKxpSbRowKUQ2fe5vCYD+nfDHqaPjaellzoaTbfVOoaKaWeAQ5nJF/hUQJh3G/Q08x31U5anGhI0IvIvxWRD4nIoSpw3i0nSGPeD5wlIscAKKWeA/4DcJ2ITCVwSpzEiKOq4QctEi1Fo6dItKSkR4dHSnjPCeGpesHRUQitGJsxScsop2cC8Avg34WnfoRAedFhnq18V02nJQ0yC/ogYkqmichjwB8ROGP+QUSmqqBWxF7gEQIv/4f1eUqp/0XgjLgRuBz4lAo88HWjSLQUjZ4i0VIvPUopFUZvvEPAq4v0956W/GnpAHo+Et73AIFj810i8jMCB/D1SqkdnUJLZiTZdsx/hHY2gmI7/0d/R1CI6PvWsV8g8CpPxnAGkZOnvUi0FI2eItHSID0Tje/He1qaR0sH0XMYI1FAB2PU+OkUWur5l0qjlyBo/2vA10TkQ8BcwnoLKrC9/XuCSnQfMk77DoEX+afAJr2VUUoNprnnaKClaPQUiZac6HnZoKehuG9Py5iiZ7MEpQL2KaV6aQBFoqURJAr68AHWEGxHNhHEUw8C54nImVCxM90Y/tP4PQK71K+Ak1Q+GW+FoaVo9BSJlqLR42kZU/Q8G9JTV6mAotLSMFJsWT4A/KHx+X8SFPy5FlgTflciSLH+HjA7/O5S4IN5bj+KREvR6CkSLUWjx9Pi6RnttDT8LCkediKBN13bqD4O/Ofw72eBPwv/Xgjc2VRiC0RL0egpEi1Fo8fT4ukZ7bQ0+i/RdKOU2quUOqBG4mAvAN4I//5j4EQReYAgUmMt1JbkzAtFoqVo9BSJlqLR42nx9Ix2WhpGhtWti2Cb8hBh2i5BJtxhBDW3j2nV6lQkWopGT5FoKRo9nhZPz2inpd5/WeLoy8A4YCdBZ/IHgL8Eykqpn6vWOhyKREvR6CkSLUWjx9Pi6RnttNSHjCvbIoKH/jnwyXauUEWipWj0FImWotHjafH0jHZa6vkn4UOkgohMB/4Q+DsVZHy1DUWipWj0FImWotHjafH0jHZa6kEmQe/h4eHhMfrQ1ubgHh4eHh7Nhxf0Hh4eHh0OL+g9PDw8Ohxe0Ht4eHh0OLyg9/Dw8OhweEHv4eHh0eHwgt7DI4SIzBaRa+o473YRuaKO867Vddw9PJoJL+g9OhIi0l3HabOBzIK+AVwLeEHv0XR4Qe8xaiFBg+bnRORXIvLPoWZ9i4isBP5WRA4RkdtE5BkR+aWIXBqeN1tEnhSRteG/94WX/DrwARF5VkS+EHYX+q8isiq8z5+G54uIfEtENorII8ARCXQuDa+xTkSWh+dfQVDe9rvh/Q5u3kh5jHX4zFiPUQkRmQf8AHifUmqniLwb+DtgKnCpUmpYghZwG5RS/0dEDgOeAU4DFEFBqv0icjxBLfGFInIu8CWl1MXhPZYARyilvioiE4CngCvDa3wWuBA4EthA0ET93gha362UejP8+5+B7ymlfiQij4f3W92EIfLwqKCe7a2HRxHwYeAepdROAKXUm2Ep8HvUSP3wjwKXiMiXws8HATOB7cC3RORUgv6fJ0Tc46ME1Qq1/X0ycDzwQYLFYRjYLiKPJtB6noj8fwSNLN4NrAd+lOlpPTwagBf0Hp2Gd4y/BbhcKbXRPEBEbgReB04hMF/uj7iWEHQRetg6/3fTEiMiBxG0oFuolNoW3vugtOd7eOQBb6P3GK14FLhSRKZAYB5xHPMw8Ge664+InBZ+Pxl4VQWNnf+QoLEEwB5gknX+Z0VkXHj+CSJyCPAzYHFow58GnBdDpxbqO0XkXYAZnWPfz8OjKfAavceohFJqvYj8J+AJERkGfuk47G+Am4DnRKQEvAxcTKBh3yci/xb4MSO7gOeAYRH5FXA78E2CSJy14WLxBvBvCHwDHyawzW8Fno6h8y0R+Q6wDngNWGX8fDtwi4jsA85WSu3LOAweHqngnbEeHh4eHQ5vuvHw8PDocHjTjYdHThCRHwDHWl9/2Xbmeni0Gt504+Hh4dHh8KYbDw8Pjw6HF/QeHh4eHQ4v6D08PDw6HF7Qe3h4eHQ4vKD38PDw6HD8/9dB5hSg/nblAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x117502cc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = trump.sample(1000).plot(x='created_at', y='intensity', style='.')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I guess, that shows what everybody knows already. The mood of the 45th President of America is fickle and can probably act as a good thing to seed your random number generator.\n",
    "\n",
    "With that saying, with so many tweets (and he generates about 8-10 tweets per day) it's too dificult to see the truth on the plot. Let's aggregate this to see the \"month average\"."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>favorite_count</th>\n",
       "      <th>id_str</th>\n",
       "      <th>in_reply_to_user_id_str</th>\n",
       "      <th>is_retweet</th>\n",
       "      <th>retweet_count</th>\n",
       "      <th>sentiment</th>\n",
       "      <th>intensity</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>month</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>2016-01</th>\n",
       "      <td>5731.091463</td>\n",
       "      <td>6.889946e+17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.026423</td>\n",
       "      <td>2152.260163</td>\n",
       "      <td>0.212208</td>\n",
       "      <td>0.246764</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-02</th>\n",
       "      <td>8208.722672</td>\n",
       "      <td>6.997745e+17</td>\n",
       "      <td>1.084716e+08</td>\n",
       "      <td>0.044534</td>\n",
       "      <td>3129.856275</td>\n",
       "      <td>0.163036</td>\n",
       "      <td>0.253002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-03</th>\n",
       "      <td>14815.274376</td>\n",
       "      <td>7.096324e+17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.047619</td>\n",
       "      <td>5428.079365</td>\n",
       "      <td>0.158371</td>\n",
       "      <td>0.272186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-04</th>\n",
       "      <td>13381.989437</td>\n",
       "      <td>7.212304e+17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.028169</td>\n",
       "      <td>4762.964789</td>\n",
       "      <td>0.316074</td>\n",
       "      <td>0.262669</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-05</th>\n",
       "      <td>15340.686610</td>\n",
       "      <td>7.321916e+17</td>\n",
       "      <td>2.657642e+08</td>\n",
       "      <td>0.019943</td>\n",
       "      <td>5254.982906</td>\n",
       "      <td>0.211409</td>\n",
       "      <td>0.270205</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         favorite_count        id_str  in_reply_to_user_id_str  is_retweet  \\\n",
       "month                                                                        \n",
       "2016-01     5731.091463  6.889946e+17                      NaN    0.026423   \n",
       "2016-02     8208.722672  6.997745e+17             1.084716e+08    0.044534   \n",
       "2016-03    14815.274376  7.096324e+17                      NaN    0.047619   \n",
       "2016-04    13381.989437  7.212304e+17                      NaN    0.028169   \n",
       "2016-05    15340.686610  7.321916e+17             2.657642e+08    0.019943   \n",
       "\n",
       "         retweet_count  sentiment  intensity  \n",
       "month                                         \n",
       "2016-01    2152.260163   0.212208   0.246764  \n",
       "2016-02    3129.856275   0.163036   0.253002  \n",
       "2016-03    5428.079365   0.158371   0.272186  \n",
       "2016-04    4762.964789   0.316074   0.262669  \n",
       "2016-05    5254.982906   0.211409   0.270205  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trump['month'] = trump.created_at.apply(lambda dt: dt.strftime('%Y-%m'))\n",
    "trump_monthly = trump.groupby('month').mean()\n",
    "trump_monthly.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x116f8def0>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1176aaac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig, axes = plt.subplots(nrows=2)\n",
    "ax0, ax1 = axes\n",
    "\n",
    "trump_monthly.sentiment.plot.bar(ax=ax0, color='k')\n",
    "trump_monthly.intensity.plot.bar(ax=ax1, color='k')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The overall sentiment of messages is positive, and the intensity is about the same throughoght two years.\n",
    "\n",
    "Then we played a bit with the retweet count to see if there's any correlation between the date and the average number of retweets, or maybe sentiment or intensity can affect that number.\n",
    "\n",
    "At this point we kind of forgot already that we wanted to find a correlation between these tweets and bitcoin prices and were driven by pure interest towards the emotions of POTUS."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x116d746a0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = trump.boxplot('retweet_count', by='month', showfliers=False)\n",
    "_ = plt.xticks(rotation=90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x116e88940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = trump.sample(1000).plot(x='intensity', y='retweet_count', style='.', logy=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1176c9470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = trump.sample(1000).plot('sentiment', 'retweet_count', style='.', logy=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looks natural that the popularity of his Twitter has grown significantly around the second half of 2016, since the moment he was officially recognized as the main Republican candidate for the White House. Other than this, nothing specifically interesting.\n",
    "\n",
    "At this point it was totally obvious that will not find any correlation between Trump tweets and Bitcoin prices, but we decided to finish this up anway. \n",
    "\n",
    "We downloaded the list of BTC prices with [Coindesk API](https://www.coindesk.com/api/) and annotated every tweet with the price Bitcoin had that day."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "import datetime\n",
    "\n",
    "\n",
    "url = 'https://api.coindesk.com/v1/bpi/historical/close.json'\n",
    "prices = requests.get(url, params={\n",
    "    'start': '2015-01-01',\n",
    "    'end': '2018-02-24',\n",
    "}).json()['bpi']\n",
    "\n",
    "trump['btc_price'] = trump.created_at.apply(lambda dt: prices[dt.strftime('%Y-%m-%d')])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then, to close down the question, we built a correlation matrix to see if there's any correlation between the values we had in our matrix."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>favorite_count</th>\n",
       "      <th>id_str</th>\n",
       "      <th>in_reply_to_user_id_str</th>\n",
       "      <th>is_retweet</th>\n",
       "      <th>retweet_count</th>\n",
       "      <th>sentiment</th>\n",
       "      <th>intensity</th>\n",
       "      <th>btc_price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>favorite_count</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.591057</td>\n",
       "      <td>-0.430882</td>\n",
       "      <td>-0.268419</td>\n",
       "      <td>0.860525</td>\n",
       "      <td>-0.012141</td>\n",
       "      <td>0.098535</td>\n",
       "      <td>0.396370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id_str</th>\n",
       "      <td>0.591057</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.567039</td>\n",
       "      <td>0.144559</td>\n",
       "      <td>0.474631</td>\n",
       "      <td>0.029834</td>\n",
       "      <td>-0.003928</td>\n",
       "      <td>0.768578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>in_reply_to_user_id_str</th>\n",
       "      <td>-0.430882</td>\n",
       "      <td>-0.567039</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.275194</td>\n",
       "      <td>-0.089284</td>\n",
       "      <td>-0.105844</td>\n",
       "      <td>-0.273243</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>is_retweet</th>\n",
       "      <td>-0.268419</td>\n",
       "      <td>0.144559</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.056263</td>\n",
       "      <td>-0.012187</td>\n",
       "      <td>-0.134873</td>\n",
       "      <td>0.097504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>retweet_count</th>\n",
       "      <td>0.860525</td>\n",
       "      <td>0.474631</td>\n",
       "      <td>-0.275194</td>\n",
       "      <td>-0.056263</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.068394</td>\n",
       "      <td>0.045699</td>\n",
       "      <td>0.312220</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sentiment</th>\n",
       "      <td>-0.012141</td>\n",
       "      <td>0.029834</td>\n",
       "      <td>-0.089284</td>\n",
       "      <td>-0.012187</td>\n",
       "      <td>-0.068394</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.231508</td>\n",
       "      <td>0.043940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>intensity</th>\n",
       "      <td>0.098535</td>\n",
       "      <td>-0.003928</td>\n",
       "      <td>-0.105844</td>\n",
       "      <td>-0.134873</td>\n",
       "      <td>0.045699</td>\n",
       "      <td>0.231508</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.014950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>btc_price</th>\n",
       "      <td>0.396370</td>\n",
       "      <td>0.768578</td>\n",
       "      <td>-0.273243</td>\n",
       "      <td>0.097504</td>\n",
       "      <td>0.312220</td>\n",
       "      <td>0.043940</td>\n",
       "      <td>0.014950</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         favorite_count    id_str  in_reply_to_user_id_str  \\\n",
       "favorite_count                 1.000000  0.591057                -0.430882   \n",
       "id_str                         0.591057  1.000000                -0.567039   \n",
       "in_reply_to_user_id_str       -0.430882 -0.567039                 1.000000   \n",
       "is_retweet                    -0.268419  0.144559                      NaN   \n",
       "retweet_count                  0.860525  0.474631                -0.275194   \n",
       "sentiment                     -0.012141  0.029834                -0.089284   \n",
       "intensity                      0.098535 -0.003928                -0.105844   \n",
       "btc_price                      0.396370  0.768578                -0.273243   \n",
       "\n",
       "                         is_retweet  retweet_count  sentiment  intensity  \\\n",
       "favorite_count            -0.268419       0.860525  -0.012141   0.098535   \n",
       "id_str                     0.144559       0.474631   0.029834  -0.003928   \n",
       "in_reply_to_user_id_str         NaN      -0.275194  -0.089284  -0.105844   \n",
       "is_retweet                 1.000000      -0.056263  -0.012187  -0.134873   \n",
       "retweet_count             -0.056263       1.000000  -0.068394   0.045699   \n",
       "sentiment                 -0.012187      -0.068394   1.000000   0.231508   \n",
       "intensity                 -0.134873       0.045699   0.231508   1.000000   \n",
       "btc_price                  0.097504       0.312220   0.043940   0.014950   \n",
       "\n",
       "                         btc_price  \n",
       "favorite_count            0.396370  \n",
       "id_str                    0.768578  \n",
       "in_reply_to_user_id_str  -0.273243  \n",
       "is_retweet                0.097504  \n",
       "retweet_count             0.312220  \n",
       "sentiment                 0.043940  \n",
       "intensity                 0.014950  \n",
       "btc_price                 1.000000  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trump.corr()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Well, as you can see, we found some interdependencies between variables. The \"0.86\" coefficient between \"favorite_count\" and \"retweet_count\" expresses the obvious fact that likes and retweets go hand in hand.\n",
    "\n",
    "As for the Bitcoin price, the closest predictor of it turned out to be a tweet ID. This doesn't come as a surprise at all. Both tweet IDs and the BTC price shows the tendency to grow steadily throughout 2016 and 2017. Needless to say, we were quite excited to finally find the main factor behind the Bitcoin price growth and decided to stop for this.\n",
    "\n",
    "\n",
    "If you like the investigation we made, you may find much more even more stunning examples, made by like-minded scientists on the [Spurious Correlations website](http://www.tylervigen.com/spurious-correlations), the website which makes you think about the oddity of the world."
   ]
  }
 ],
 "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}