datascrum-altdata-sample-nb-v1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Alternative Data exploration for Revenue prediction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Load dependency libraries "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from textblob import TextBlob\n",
    "import re\n",
    "import matplotlib.pyplot as plt\n",
    "plt.style.use('seaborn')\n",
    "%matplotlib inline\n",
    "from pandas.io.json import json_normalize\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Use EODHistorical API provided to source Revenue data for BKNG.US under the Financial -> Income_Statement -> Quarterly  data dictionary keys\n",
    "\n",
    "The API supports fields filtering with the parameter ‘filter=’. With ability to specify a block, field or Code required within the json data. Also different layers can be divided with “::” and it’s possible to have any number of layers. In our case, we want the quarterly Income_Statement and hence have configured API accordingly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Specify API url link to for quarterly income statement retrival\n",
    "eodhist_url = 'http://eodhistoricaldata.com/api/fundamentals/BKNG.US?api_token=5cc0ea63d1cda3.37070012&filter=Financials::Income_Statement::quarterly'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Use a combination of pandas json reader to read json data as series\n",
    "#And normalise with json normaliser and finally retrive the required column fields\n",
    "fund_df = json_normalize(pd.read_json(eodhist_url,typ='series'))[['date','totalRevenue','costOfRevenue']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "         date   totalRevenue costOfRevenue\n",
      "0  2019-03-31  2837000000.00  501000000.00\n",
      "1  2018-12-31  3212615000.00  242000000.00\n",
      "2  2018-09-30  4849090000.00          0.00\n",
      "3  2018-06-30  3537094000.00          0.00\n",
      "4  2018-03-31  2928201000.00          0.00\n",
      "\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 77 entries, 0 to 76\n",
      "Data columns (total 3 columns):\n",
      "date             77 non-null object\n",
      "totalRevenue     77 non-null object\n",
      "costOfRevenue    77 non-null object\n",
      "dtypes: object(3)\n",
      "memory usage: 1.9+ KB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "#Explore the data\n",
    "print(fund_df.head())\n",
    "print()\n",
    "print(fund_df.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#set date as index for later use in dataframe join operation\n",
    "fund_df.index = pd.to_datetime(fund_df.date)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Load tweet data extracted from altcap database for BKNG.US into a dataframe and process data for senitments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load tweet data from filesystem\n",
    "tweet_df = pd.read_csv('C:\\\\Users\\\\eche1\\\\Downloads\\\\tweets_bkng.csv', parse_dates = ['date'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# define function to be used for tweet senitments analysis \n",
    "def clean_get_twt_sentiment(tweet): \n",
    "        ''' \n",
    "        Utility function to clean tweet text by removing links, special characters \n",
    "        using simple regex statements and to classify sentiment of passed tweet \n",
    "        using textblob's sentiment method \n",
    "        '''\n",
    "        #clean tweet\n",
    "        clean_tweet = ' '.join(re.sub(\"(@[A-Za-z0-9]+)|([^0-9A-Za-z \\t])|(\\w+:\\/\\/\\S+)\", \" \", tweet).split())\n",
    "        # create TextBlob object of passed tweet text \n",
    "        analysis = TextBlob(clean_tweet) \n",
    "        # set sentiment \n",
    "        if analysis.sentiment.polarity > 0: \n",
    "            return 'positive'\n",
    "        elif analysis.sentiment.polarity == 0: \n",
    "            return 'neutral'\n",
    "        else: \n",
    "            return 'negative'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>text</th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>ticker</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2015-04-09 08:45:00</td>\n",
       "      <td>Q5. What activity organised by an accommodatio...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>BKNG.O</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2015-04-09 08:45:00</td>\n",
       "      <td>Q5. What activity organised by an accommodatio...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>BKNG.O</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2015-04-09 08:51:00</td>\n",
       "      <td>Q6 #BookingsBest lists are based on reviews wi...</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>BKNG.O</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2015-04-09 08:52:00</td>\n",
       "      <td>#BookingsBest lists are based on reviews with ...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>BKNG.O</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2015-04-09 08:52:00</td>\n",
       "      <td>#BookingsBest lists are based on reviews with ...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>BKNG.O</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 date                                               text  \\\n",
       "0 2015-04-09 08:45:00  Q5. What activity organised by an accommodatio...   \n",
       "1 2015-04-09 08:45:00  Q5. What activity organised by an accommodatio...   \n",
       "2 2015-04-09 08:51:00  Q6 #BookingsBest lists are based on reviews wi...   \n",
       "3 2015-04-09 08:52:00  #BookingsBest lists are based on reviews with ...   \n",
       "4 2015-04-09 08:52:00  #BookingsBest lists are based on reviews with ...   \n",
       "\n",
       "   likes  retweet  count_comments  ticker  \n",
       "0      0        0               0  BKNG.O  \n",
       "1      0        0               0  BKNG.O  \n",
       "2      3        1               5  BKNG.O  \n",
       "3      1        1               0  BKNG.O  \n",
       "4      0        0               0  BKNG.O  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore tweet data \n",
    "tweet_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Process the tweet text for senitment indicator\n",
    "tweet_df['senti']  = tweet_df['text'].apply(clean_get_twt_sentiment)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>date</th>\n",
       "      <th>text</th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>ticker</th>\n",
       "      <th>senti</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2015-04-09 08:45:00</td>\n",
       "      <td>Q5. What activity organised by an accommodatio...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>BKNG.O</td>\n",
       "      <td>neutral</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2015-04-09 08:45:00</td>\n",
       "      <td>Q5. What activity organised by an accommodatio...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>BKNG.O</td>\n",
       "      <td>neutral</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2015-04-09 08:51:00</td>\n",
       "      <td>Q6 #BookingsBest lists are based on reviews wi...</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>BKNG.O</td>\n",
       "      <td>positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2015-04-09 08:52:00</td>\n",
       "      <td>#BookingsBest lists are based on reviews with ...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>BKNG.O</td>\n",
       "      <td>positive</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2015-04-09 08:52:00</td>\n",
       "      <td>#BookingsBest lists are based on reviews with ...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>BKNG.O</td>\n",
       "      <td>positive</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 date                                               text  \\\n",
       "0 2015-04-09 08:45:00  Q5. What activity organised by an accommodatio...   \n",
       "1 2015-04-09 08:45:00  Q5. What activity organised by an accommodatio...   \n",
       "2 2015-04-09 08:51:00  Q6 #BookingsBest lists are based on reviews wi...   \n",
       "3 2015-04-09 08:52:00  #BookingsBest lists are based on reviews with ...   \n",
       "4 2015-04-09 08:52:00  #BookingsBest lists are based on reviews with ...   \n",
       "\n",
       "   likes  retweet  count_comments  ticker     senti  \n",
       "0      0        0               0  BKNG.O   neutral  \n",
       "1      0        0               0  BKNG.O   neutral  \n",
       "2      3        1               5  BKNG.O  positive  \n",
       "3      1        1               0  BKNG.O  positive  \n",
       "4      0        0               0  BKNG.O  positive  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore tweet data for senitment indicator\n",
    "tweet_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#set date as index for later use in dataframe time series, resampling (quarterly) and  join operations\n",
    "tweet_df.index = pd.to_datetime(tweet_df.date)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "#select required features columns and convert categorical senitment into indicator variables\n",
    "tweet_df2 = pd.get_dummies(tweet_df[['likes','retweet','count_comments','senti']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "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>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</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>2015-04-09 08:45:00</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-09 08:45:00</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-09 08:51:00</th>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-09 08:52:00</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-09 08:52:00</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     likes  retweet  count_comments  senti_negative  \\\n",
       "date                                                                  \n",
       "2015-04-09 08:45:00      0        0               0               0   \n",
       "2015-04-09 08:45:00      0        0               0               0   \n",
       "2015-04-09 08:51:00      3        1               5               0   \n",
       "2015-04-09 08:52:00      1        1               0               0   \n",
       "2015-04-09 08:52:00      0        0               0               0   \n",
       "\n",
       "                     senti_neutral  senti_positive  \n",
       "date                                                \n",
       "2015-04-09 08:45:00              1               0  \n",
       "2015-04-09 08:45:00              1               0  \n",
       "2015-04-09 08:51:00              0               1  \n",
       "2015-04-09 08:52:00              0               1  \n",
       "2015-04-09 08:52:00              0               1  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore new data frame\n",
    "tweet_df2.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-04-09</th>\n",
       "      <td>11</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-10</th>\n",
       "      <td>20</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-11</th>\n",
       "      <td>16</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-12</th>\n",
       "      <td>14</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-13</th>\n",
       "      <td>11</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            likes  retweet  count_comments  senti_negative  senti_neutral  \\\n",
       "2015-04-09     11        5               9               0              2   \n",
       "2015-04-10     20        2               0               0              0   \n",
       "2015-04-11     16        2               0               0              1   \n",
       "2015-04-12     14        5               1               0              1   \n",
       "2015-04-13     11        4               1               0              1   \n",
       "\n",
       "            senti_positive  \n",
       "2015-04-09               8  \n",
       "2015-04-10               1  \n",
       "2015-04-11               0  \n",
       "2015-04-12               0  \n",
       "2015-04-13               0  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore summing the individual column by date\n",
    "tweet_df2.groupby([tweet_df2.index.date]).aggregate(np.sum).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Sum tweet feature values by date\n",
    "tweet_sum_df = tweet_df2.groupby([tweet_df2.index.date]).aggregate(np.sum)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 533 entries, 2015-04-09 to 2019-05-16\n",
      "Data columns (total 6 columns):\n",
      "likes             533 non-null int64\n",
      "retweet           533 non-null int64\n",
      "count_comments    533 non-null int64\n",
      "senti_negative    533 non-null uint8\n",
      "senti_neutral     533 non-null uint8\n",
      "senti_positive    533 non-null uint8\n",
      "dtypes: int64(3), uint8(3)\n",
      "memory usage: 18.2+ KB\n"
     ]
    }
   ],
   "source": [
    "tweet_sum_df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "#reset date as index for later use in dataframe time series, resampling (quarterly) and  join operations\n",
    "tweet_sum_df.index = pd.to_datetime(tweet_sum_df.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-06-30</th>\n",
       "      <td>896</td>\n",
       "      <td>348</td>\n",
       "      <td>152</td>\n",
       "      <td>4</td>\n",
       "      <td>39</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-09-30</th>\n",
       "      <td>263</td>\n",
       "      <td>190</td>\n",
       "      <td>139</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-12-31</th>\n",
       "      <td>154</td>\n",
       "      <td>57</td>\n",
       "      <td>35</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-03-31</th>\n",
       "      <td>171</td>\n",
       "      <td>85</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-06-30</th>\n",
       "      <td>585</td>\n",
       "      <td>141</td>\n",
       "      <td>64</td>\n",
       "      <td>3</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-09-30</th>\n",
       "      <td>1051</td>\n",
       "      <td>327</td>\n",
       "      <td>278</td>\n",
       "      <td>3</td>\n",
       "      <td>16</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31</th>\n",
       "      <td>556</td>\n",
       "      <td>120</td>\n",
       "      <td>45</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-31</th>\n",
       "      <td>494</td>\n",
       "      <td>130</td>\n",
       "      <td>64</td>\n",
       "      <td>4</td>\n",
       "      <td>7</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-06-30</th>\n",
       "      <td>742</td>\n",
       "      <td>183</td>\n",
       "      <td>313</td>\n",
       "      <td>2</td>\n",
       "      <td>13</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-09-30</th>\n",
       "      <td>1160</td>\n",
       "      <td>250</td>\n",
       "      <td>679</td>\n",
       "      <td>9</td>\n",
       "      <td>32</td>\n",
       "      <td>82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31</th>\n",
       "      <td>1492</td>\n",
       "      <td>303</td>\n",
       "      <td>500</td>\n",
       "      <td>8</td>\n",
       "      <td>34</td>\n",
       "      <td>105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-03-31</th>\n",
       "      <td>1199</td>\n",
       "      <td>159</td>\n",
       "      <td>222</td>\n",
       "      <td>9</td>\n",
       "      <td>27</td>\n",
       "      <td>62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-30</th>\n",
       "      <td>897</td>\n",
       "      <td>112</td>\n",
       "      <td>217</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-30</th>\n",
       "      <td>344</td>\n",
       "      <td>68</td>\n",
       "      <td>385</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-12-31</th>\n",
       "      <td>246</td>\n",
       "      <td>66</td>\n",
       "      <td>260</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-03-31</th>\n",
       "      <td>302</td>\n",
       "      <td>53</td>\n",
       "      <td>285</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2019-06-30</th>\n",
       "      <td>202</td>\n",
       "      <td>28</td>\n",
       "      <td>205</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            likes  retweet  count_comments  senti_negative  senti_neutral  \\\n",
       "2015-06-30    896      348             152               4             39   \n",
       "2015-09-30    263      190             139               0             10   \n",
       "2015-12-31    154       57              35               1              9   \n",
       "2016-03-31    171       85              15               0              3   \n",
       "2016-06-30    585      141              64               3              6   \n",
       "2016-09-30   1051      327             278               3             16   \n",
       "2016-12-31    556      120              45               2              7   \n",
       "2017-03-31    494      130              64               4              7   \n",
       "2017-06-30    742      183             313               2             13   \n",
       "2017-09-30   1160      250             679               9             32   \n",
       "2017-12-31   1492      303             500               8             34   \n",
       "2018-03-31   1199      159             222               9             27   \n",
       "2018-06-30    897      112             217               5              5   \n",
       "2018-09-30    344       68             385               0              2   \n",
       "2018-12-31    246       66             260               0              2   \n",
       "2019-03-31    302       53             285               2              0   \n",
       "2019-06-30    202       28             205               1              1   \n",
       "\n",
       "            senti_positive  \n",
       "2015-06-30              48  \n",
       "2015-09-30              17  \n",
       "2015-12-31               9  \n",
       "2016-03-31               3  \n",
       "2016-06-30              12  \n",
       "2016-09-30              27  \n",
       "2016-12-31              14  \n",
       "2017-03-31              20  \n",
       "2017-06-30              47  \n",
       "2017-09-30              82  \n",
       "2017-12-31             105  \n",
       "2018-03-31              62  \n",
       "2018-06-30              50  \n",
       "2018-09-30               8  \n",
       "2018-12-31               8  \n",
       "2019-03-31              11  \n",
       "2019-06-30               9  "
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore the resampling by Quarterly and aggregation.. this should be similar to cumsun function\n",
    "tweet_sum_df.resample('Q', label='right').sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "likes             896\n",
      "retweet           348\n",
      "count_comments    152\n",
      "senti_negative      4\n",
      "senti_neutral      39\n",
      "senti_positive     48\n",
      "dtype: int64\n",
      "\n",
      "likes             263\n",
      "retweet           190\n",
      "count_comments    139\n",
      "senti_negative      0\n",
      "senti_neutral      10\n",
      "senti_positive     17\n",
      "dtype: int64\n"
     ]
    }
   ],
   "source": [
    "# Check the Resampling and aggreggation is correct... using specific quarter date.\n",
    "print(tweet_sum_df.loc[:'2015-06-30'].sum())\n",
    "print()\n",
    "print(tweet_sum_df.loc['2015-06-30': '2015-09-30'].sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Resample tweet data to Quarterly to match Revenue time series\n",
    "tweet_sum_quarterly_df = tweet_sum_df.resample('Q', label='right').sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "DatetimeIndex: 17 entries, 2015-06-30 to 2019-06-30\n",
      "Freq: Q-DEC\n",
      "Data columns (total 6 columns):\n",
      "likes             17 non-null int64\n",
      "retweet           17 non-null int64\n",
      "count_comments    17 non-null int64\n",
      "senti_negative    17 non-null uint8\n",
      "senti_neutral     17 non-null uint8\n",
      "senti_positive    17 non-null uint8\n",
      "dtypes: int64(3), uint8(3)\n",
      "memory usage: 595.0 bytes\n"
     ]
    }
   ],
   "source": [
    "#Explore data\n",
    "tweet_sum_quarterly_df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>totalRevenue</th>\n",
       "      <th>costOfRevenue</th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-03-31</th>\n",
       "      <td>2837000000.00</td>\n",
       "      <td>501000000.00</td>\n",
       "      <td>302</td>\n",
       "      <td>53</td>\n",
       "      <td>285</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-12-31</th>\n",
       "      <td>3212615000.00</td>\n",
       "      <td>242000000.00</td>\n",
       "      <td>246</td>\n",
       "      <td>66</td>\n",
       "      <td>260</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-30</th>\n",
       "      <td>4849090000.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>344</td>\n",
       "      <td>68</td>\n",
       "      <td>385</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-30</th>\n",
       "      <td>3537094000.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>897</td>\n",
       "      <td>112</td>\n",
       "      <td>217</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-03-31</th>\n",
       "      <td>2928201000.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1199</td>\n",
       "      <td>159</td>\n",
       "      <td>222</td>\n",
       "      <td>9</td>\n",
       "      <td>27</td>\n",
       "      <td>62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31</th>\n",
       "      <td>2803093000.00</td>\n",
       "      <td>45150000.00</td>\n",
       "      <td>1492</td>\n",
       "      <td>303</td>\n",
       "      <td>500</td>\n",
       "      <td>8</td>\n",
       "      <td>34</td>\n",
       "      <td>105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-09-30</th>\n",
       "      <td>4434029000.00</td>\n",
       "      <td>54181000.00</td>\n",
       "      <td>1160</td>\n",
       "      <td>250</td>\n",
       "      <td>679</td>\n",
       "      <td>9</td>\n",
       "      <td>32</td>\n",
       "      <td>82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-06-30</th>\n",
       "      <td>3024556000.00</td>\n",
       "      <td>67425000.00</td>\n",
       "      <td>742</td>\n",
       "      <td>183</td>\n",
       "      <td>313</td>\n",
       "      <td>2</td>\n",
       "      <td>13</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-31</th>\n",
       "      <td>2419404000.00</td>\n",
       "      <td>80401000.00</td>\n",
       "      <td>494</td>\n",
       "      <td>130</td>\n",
       "      <td>64</td>\n",
       "      <td>4</td>\n",
       "      <td>7</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31</th>\n",
       "      <td>2348433000.00</td>\n",
       "      <td>72072000.00</td>\n",
       "      <td>556</td>\n",
       "      <td>120</td>\n",
       "      <td>45</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-09-30</th>\n",
       "      <td>3690552000.00</td>\n",
       "      <td>101489000.00</td>\n",
       "      <td>1051</td>\n",
       "      <td>327</td>\n",
       "      <td>278</td>\n",
       "      <td>3</td>\n",
       "      <td>16</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-06-30</th>\n",
       "      <td>2555902000.00</td>\n",
       "      <td>126084000.00</td>\n",
       "      <td>585</td>\n",
       "      <td>141</td>\n",
       "      <td>64</td>\n",
       "      <td>3</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-03-31</th>\n",
       "      <td>2148119000.00</td>\n",
       "      <td>128669000.00</td>\n",
       "      <td>171</td>\n",
       "      <td>85</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-12-31</th>\n",
       "      <td>1999995000.00</td>\n",
       "      <td>120612000.00</td>\n",
       "      <td>154</td>\n",
       "      <td>57</td>\n",
       "      <td>35</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-09-30</th>\n",
       "      <td>3102901000.00</td>\n",
       "      <td>169274000.00</td>\n",
       "      <td>263</td>\n",
       "      <td>190</td>\n",
       "      <td>139</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-06-30</th>\n",
       "      <td>2280397000.00</td>\n",
       "      <td>187491000.00</td>\n",
       "      <td>896</td>\n",
       "      <td>348</td>\n",
       "      <td>152</td>\n",
       "      <td>4</td>\n",
       "      <td>39</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             totalRevenue costOfRevenue  likes  retweet  count_comments  \\\n",
       "2019-03-31  2837000000.00  501000000.00    302       53             285   \n",
       "2018-12-31  3212615000.00  242000000.00    246       66             260   \n",
       "2018-09-30  4849090000.00          0.00    344       68             385   \n",
       "2018-06-30  3537094000.00          0.00    897      112             217   \n",
       "2018-03-31  2928201000.00          0.00   1199      159             222   \n",
       "2017-12-31  2803093000.00   45150000.00   1492      303             500   \n",
       "2017-09-30  4434029000.00   54181000.00   1160      250             679   \n",
       "2017-06-30  3024556000.00   67425000.00    742      183             313   \n",
       "2017-03-31  2419404000.00   80401000.00    494      130              64   \n",
       "2016-12-31  2348433000.00   72072000.00    556      120              45   \n",
       "2016-09-30  3690552000.00  101489000.00   1051      327             278   \n",
       "2016-06-30  2555902000.00  126084000.00    585      141              64   \n",
       "2016-03-31  2148119000.00  128669000.00    171       85              15   \n",
       "2015-12-31  1999995000.00  120612000.00    154       57              35   \n",
       "2015-09-30  3102901000.00  169274000.00    263      190             139   \n",
       "2015-06-30  2280397000.00  187491000.00    896      348             152   \n",
       "\n",
       "            senti_negative  senti_neutral  senti_positive  \n",
       "2019-03-31               2              0              11  \n",
       "2018-12-31               0              2               8  \n",
       "2018-09-30               0              2               8  \n",
       "2018-06-30               5              5              50  \n",
       "2018-03-31               9             27              62  \n",
       "2017-12-31               8             34             105  \n",
       "2017-09-30               9             32              82  \n",
       "2017-06-30               2             13              47  \n",
       "2017-03-31               4              7              20  \n",
       "2016-12-31               2              7              14  \n",
       "2016-09-30               3             16              27  \n",
       "2016-06-30               3              6              12  \n",
       "2016-03-31               0              3               3  \n",
       "2015-12-31               1              9               9  \n",
       "2015-09-30               0             10              17  \n",
       "2015-06-30               4             39              48  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Join tweet and revenue dataframe by quarterly date\n",
    "tweet_fund_df = fund_df[['totalRevenue','costOfRevenue']].join(tweet_sum_quarterly_df, how='inner')\n",
    "tweet_fund_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Drop costOfRevenue as this will not be needed\n",
    "tweet_fund_df.drop('costOfRevenue', axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:625: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\sklearn\\base.py:462: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    }
   ],
   "source": [
    "#Normalise and standardise the data values\n",
    "from sklearn import preprocessing as prep\n",
    "#tweet_fund_df_nom = pd.DataFrame(prep.MinMaxScaler(feature_range=(0, 1)).fit_transform(tweet_fund_df), columns=tweet_fund_df.columns, index=tweet_fund_df.index)\n",
    "tweet_fund_df_nom = pd.DataFrame(prep.StandardScaler().fit_transform(tweet_fund_df), columns=tweet_fund_df.columns, index=tweet_fund_df.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>totalRevenue</th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-03-31</th>\n",
       "      <td>-0.224631</td>\n",
       "      <td>-0.886782</td>\n",
       "      <td>-1.151780</td>\n",
       "      <td>0.320742</td>\n",
       "      <td>-0.415227</td>\n",
       "      <td>-1.081182</td>\n",
       "      <td>-0.743672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-12-31</th>\n",
       "      <td>0.261087</td>\n",
       "      <td>-1.025690</td>\n",
       "      <td>-1.014412</td>\n",
       "      <td>0.179290</td>\n",
       "      <td>-1.079591</td>\n",
       "      <td>-0.917985</td>\n",
       "      <td>-0.846543</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-30</th>\n",
       "      <td>2.377258</td>\n",
       "      <td>-0.782600</td>\n",
       "      <td>-0.993278</td>\n",
       "      <td>0.886550</td>\n",
       "      <td>-1.079591</td>\n",
       "      <td>-0.917985</td>\n",
       "      <td>-0.846543</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-30</th>\n",
       "      <td>0.680680</td>\n",
       "      <td>0.589121</td>\n",
       "      <td>-0.528339</td>\n",
       "      <td>-0.064007</td>\n",
       "      <td>0.581318</td>\n",
       "      <td>-0.673189</td>\n",
       "      <td>0.593652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-03-31</th>\n",
       "      <td>-0.106696</td>\n",
       "      <td>1.338234</td>\n",
       "      <td>-0.031700</td>\n",
       "      <td>-0.035717</td>\n",
       "      <td>1.910046</td>\n",
       "      <td>1.121982</td>\n",
       "      <td>1.005136</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31</th>\n",
       "      <td>-0.268477</td>\n",
       "      <td>2.065023</td>\n",
       "      <td>1.489917</td>\n",
       "      <td>1.537230</td>\n",
       "      <td>1.577864</td>\n",
       "      <td>1.693173</td>\n",
       "      <td>2.479620</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-09-30</th>\n",
       "      <td>1.840531</td>\n",
       "      <td>1.241495</td>\n",
       "      <td>0.929877</td>\n",
       "      <td>2.550026</td>\n",
       "      <td>1.910046</td>\n",
       "      <td>1.529975</td>\n",
       "      <td>1.690942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-06-30</th>\n",
       "      <td>0.017903</td>\n",
       "      <td>0.204642</td>\n",
       "      <td>0.221903</td>\n",
       "      <td>0.479169</td>\n",
       "      <td>-0.415227</td>\n",
       "      <td>-0.020400</td>\n",
       "      <td>0.490781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-31</th>\n",
       "      <td>-0.764636</td>\n",
       "      <td>-0.410524</td>\n",
       "      <td>-0.338137</td>\n",
       "      <td>-0.929693</td>\n",
       "      <td>0.249136</td>\n",
       "      <td>-0.509992</td>\n",
       "      <td>-0.435059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31</th>\n",
       "      <td>-0.856410</td>\n",
       "      <td>-0.256733</td>\n",
       "      <td>-0.443805</td>\n",
       "      <td>-1.037197</td>\n",
       "      <td>-0.415227</td>\n",
       "      <td>-0.509992</td>\n",
       "      <td>-0.640801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-09-30</th>\n",
       "      <td>0.879121</td>\n",
       "      <td>0.971119</td>\n",
       "      <td>1.743520</td>\n",
       "      <td>0.281136</td>\n",
       "      <td>-0.083045</td>\n",
       "      <td>0.224396</td>\n",
       "      <td>-0.195026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-06-30</th>\n",
       "      <td>-0.588126</td>\n",
       "      <td>-0.184798</td>\n",
       "      <td>-0.221903</td>\n",
       "      <td>-0.929693</td>\n",
       "      <td>-0.083045</td>\n",
       "      <td>-0.591590</td>\n",
       "      <td>-0.709381</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-03-31</th>\n",
       "      <td>-1.115442</td>\n",
       "      <td>-1.211728</td>\n",
       "      <td>-0.813643</td>\n",
       "      <td>-1.206939</td>\n",
       "      <td>-1.079591</td>\n",
       "      <td>-0.836386</td>\n",
       "      <td>-1.017995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-12-31</th>\n",
       "      <td>-1.306985</td>\n",
       "      <td>-1.253897</td>\n",
       "      <td>-1.109513</td>\n",
       "      <td>-1.093778</td>\n",
       "      <td>-0.747409</td>\n",
       "      <td>-0.346794</td>\n",
       "      <td>-0.812252</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-09-30</th>\n",
       "      <td>0.119213</td>\n",
       "      <td>-0.983522</td>\n",
       "      <td>0.295870</td>\n",
       "      <td>-0.505337</td>\n",
       "      <td>-1.079591</td>\n",
       "      <td>-0.265196</td>\n",
       "      <td>-0.537930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-06-30</th>\n",
       "      <td>-0.944389</td>\n",
       "      <td>0.586640</td>\n",
       "      <td>1.965422</td>\n",
       "      <td>-0.431782</td>\n",
       "      <td>0.249136</td>\n",
       "      <td>2.101166</td>\n",
       "      <td>0.525071</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            totalRevenue     likes   retweet  count_comments  senti_negative  \\\n",
       "2019-03-31     -0.224631 -0.886782 -1.151780        0.320742       -0.415227   \n",
       "2018-12-31      0.261087 -1.025690 -1.014412        0.179290       -1.079591   \n",
       "2018-09-30      2.377258 -0.782600 -0.993278        0.886550       -1.079591   \n",
       "2018-06-30      0.680680  0.589121 -0.528339       -0.064007        0.581318   \n",
       "2018-03-31     -0.106696  1.338234 -0.031700       -0.035717        1.910046   \n",
       "2017-12-31     -0.268477  2.065023  1.489917        1.537230        1.577864   \n",
       "2017-09-30      1.840531  1.241495  0.929877        2.550026        1.910046   \n",
       "2017-06-30      0.017903  0.204642  0.221903        0.479169       -0.415227   \n",
       "2017-03-31     -0.764636 -0.410524 -0.338137       -0.929693        0.249136   \n",
       "2016-12-31     -0.856410 -0.256733 -0.443805       -1.037197       -0.415227   \n",
       "2016-09-30      0.879121  0.971119  1.743520        0.281136       -0.083045   \n",
       "2016-06-30     -0.588126 -0.184798 -0.221903       -0.929693       -0.083045   \n",
       "2016-03-31     -1.115442 -1.211728 -0.813643       -1.206939       -1.079591   \n",
       "2015-12-31     -1.306985 -1.253897 -1.109513       -1.093778       -0.747409   \n",
       "2015-09-30      0.119213 -0.983522  0.295870       -0.505337       -1.079591   \n",
       "2015-06-30     -0.944389  0.586640  1.965422       -0.431782        0.249136   \n",
       "\n",
       "            senti_neutral  senti_positive  \n",
       "2019-03-31      -1.081182       -0.743672  \n",
       "2018-12-31      -0.917985       -0.846543  \n",
       "2018-09-30      -0.917985       -0.846543  \n",
       "2018-06-30      -0.673189        0.593652  \n",
       "2018-03-31       1.121982        1.005136  \n",
       "2017-12-31       1.693173        2.479620  \n",
       "2017-09-30       1.529975        1.690942  \n",
       "2017-06-30      -0.020400        0.490781  \n",
       "2017-03-31      -0.509992       -0.435059  \n",
       "2016-12-31      -0.509992       -0.640801  \n",
       "2016-09-30       0.224396       -0.195026  \n",
       "2016-06-30      -0.591590       -0.709381  \n",
       "2016-03-31      -0.836386       -1.017995  \n",
       "2015-12-31      -0.346794       -0.812252  \n",
       "2015-09-30      -0.265196       -0.537930  \n",
       "2015-06-30       2.101166        0.525071  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore normalised dataframe\n",
    "tweet_fund_df_nom"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "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>totalRevenue</th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>totalRevenue</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.242425</td>\n",
       "      <td>0.089407</td>\n",
       "      <td>0.740516</td>\n",
       "      <td>0.154957</td>\n",
       "      <td>0.021636</td>\n",
       "      <td>0.215627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>likes</th>\n",
       "      <td>0.242425</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.749258</td>\n",
       "      <td>0.585070</td>\n",
       "      <td>0.884749</td>\n",
       "      <td>0.795292</td>\n",
       "      <td>0.911202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>retweet</th>\n",
       "      <td>0.089407</td>\n",
       "      <td>0.749258</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.377470</td>\n",
       "      <td>0.510280</td>\n",
       "      <td>0.836692</td>\n",
       "      <td>0.652346</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>count_comments</th>\n",
       "      <td>0.740516</td>\n",
       "      <td>0.585070</td>\n",
       "      <td>0.377470</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.532460</td>\n",
       "      <td>0.462752</td>\n",
       "      <td>0.686259</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>senti_negative</th>\n",
       "      <td>0.154957</td>\n",
       "      <td>0.884749</td>\n",
       "      <td>0.510280</td>\n",
       "      <td>0.532460</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.742016</td>\n",
       "      <td>0.874409</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>senti_neutral</th>\n",
       "      <td>0.021636</td>\n",
       "      <td>0.795292</td>\n",
       "      <td>0.836692</td>\n",
       "      <td>0.462752</td>\n",
       "      <td>0.742016</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.822844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>senti_positive</th>\n",
       "      <td>0.215627</td>\n",
       "      <td>0.911202</td>\n",
       "      <td>0.652346</td>\n",
       "      <td>0.686259</td>\n",
       "      <td>0.874409</td>\n",
       "      <td>0.822844</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                totalRevenue     likes   retweet  count_comments  \\\n",
       "totalRevenue        1.000000  0.242425  0.089407        0.740516   \n",
       "likes               0.242425  1.000000  0.749258        0.585070   \n",
       "retweet             0.089407  0.749258  1.000000        0.377470   \n",
       "count_comments      0.740516  0.585070  0.377470        1.000000   \n",
       "senti_negative      0.154957  0.884749  0.510280        0.532460   \n",
       "senti_neutral       0.021636  0.795292  0.836692        0.462752   \n",
       "senti_positive      0.215627  0.911202  0.652346        0.686259   \n",
       "\n",
       "                senti_negative  senti_neutral  senti_positive  \n",
       "totalRevenue          0.154957       0.021636        0.215627  \n",
       "likes                 0.884749       0.795292        0.911202  \n",
       "retweet               0.510280       0.836692        0.652346  \n",
       "count_comments        0.532460       0.462752        0.686259  \n",
       "senti_negative        1.000000       0.742016        0.874409  \n",
       "senti_neutral         0.742016       1.000000        0.822844  \n",
       "senti_positive        0.874409       0.822844        1.000000  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore correlation between Revenue and other feature values\n",
    "tweet_fund_df_nom.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x17b91f85160>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Explore visual correlation between Revenue and other feature values\n",
    "sns.heatmap(tweet_fund_df_nom.corr(), annot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([<matplotlib.axes._subplots.AxesSubplot object at 0x0000017B91F2D438>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B91F16C50>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B91F038D0>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B91EF14A8>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B91EDB1D0>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B91EC6DD8>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B91EB1A20>],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 7 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Explore visual correlation between Revenue and other feature values\n",
    "tweet_fund_df_nom.plot(subplots=True, figsize=(10,6))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'likes_0': 0.24242485925361343},\n",
       " {'likes_1': 0.1377245548605161},\n",
       " {'likes_2': -0.09931723995146831},\n",
       " {'likes_3': -0.08461449533731318},\n",
       " {'likes_4': -0.10675441759345405},\n",
       " {'retweet_0': 0.08940743062571363},\n",
       " {'retweet_1': -0.06768196380798862},\n",
       " {'retweet_2': -0.356740722273298},\n",
       " {'retweet_3': -0.40819912718815954},\n",
       " {'retweet_4': -0.07544432239596847},\n",
       " {'count_comments_0': 0.7405158540854997},\n",
       " {'count_comments_1': 0.39766299851108655},\n",
       " {'count_comments_2': -0.00035689765884525706},\n",
       " {'count_comments_3': -0.03011492658646072},\n",
       " {'count_comments_4': 0.6120657017565297},\n",
       " {'senti_negative_0': 0.1549570044879294},\n",
       " {'senti_negative_1': 0.07399691940041826},\n",
       " {'senti_negative_2': 0.12554416031877225},\n",
       " {'senti_negative_3': -0.02160128308349073},\n",
       " {'senti_negative_4': -0.14305887708002446},\n",
       " {'senti_neutral_0': 0.021635929529872634},\n",
       " {'senti_neutral_1': 0.1259903230858425},\n",
       " {'senti_neutral_2': -0.1028306947411578},\n",
       " {'senti_neutral_3': -0.30682796770003956},\n",
       " {'senti_neutral_4': -0.06527303203608649},\n",
       " {'senti_positive_0': 0.2156274732281517},\n",
       " {'senti_positive_1': 0.22542981834402565},\n",
       " {'senti_positive_2': -0.028564438375690938},\n",
       " {'senti_positive_3': 0.02207305422646849},\n",
       " {'senti_positive_4': -0.044632623569311344}]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#checking whether more correleration can be observed from a time shift or lag of the features with features \n",
    "#that were not originally correlated showing good signal.\n",
    "\n",
    "feats = tweet_fund_df_nom.drop('totalRevenue', axis=1).columns.values\n",
    "[ {x + '_' + str(y) : tweet_fund_df_nom[x].shift(y).corr(tweet_fund_df_nom.totalRevenue)} for x in feats for y in range(5) ]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Prepare data for using in Stats and Machine Learning Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((16, 6), (16,))"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Select target(dependent/label) and feature(independent) variables\n",
    "targets = tweet_fund_df_nom['totalRevenue']\n",
    "features = tweet_fund_df_nom.drop(['totalRevenue'], axis=1)\n",
    "feature_names = features.columns\n",
    "feature_names = features.columns\n",
    "features.shape, targets.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Using a linear OLS model here \n",
    "import statsmodels.api as sm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add a constant to the features\n",
    "linear_features = sm.add_constant(features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['const', 'likes', 'retweet', 'count_comments', 'senti_negative',\n",
       "       'senti_neutral', 'senti_positive'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore features columns\n",
    "linear_features.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(16, 7) (13, 7) (3, 7)\n"
     ]
    }
   ],
   "source": [
    "# Create a size for the training set that is 85% of the total number of samples\n",
    "train_size = int(0.85 * targets.shape[0])\n",
    "train_features = linear_features[:train_size]\n",
    "train_targets = targets[:train_size]\n",
    "test_features = linear_features[train_size:]\n",
    "test_targets = targets[train_size:]\n",
    "print(linear_features.shape, train_features.shape, test_features.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:           totalRevenue   R-squared:                       0.825\n",
      "Model:                            OLS   Adj. R-squared:                  0.649\n",
      "Method:                 Least Squares   F-statistic:                     4.701\n",
      "Date:                Mon, 17 Jun 2019   Prob (F-statistic):             0.0408\n",
      "Time:                        00:40:04   Log-Likelihood:                -7.1495\n",
      "No. Observations:                  13   AIC:                             28.30\n",
      "Df Residuals:                       6   BIC:                             32.25\n",
      "Df Model:                           6                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==================================================================================\n",
      "                     coef    std err          t      P>|t|      [0.025      0.975]\n",
      "----------------------------------------------------------------------------------\n",
      "const             -0.2129      0.217     -0.979      0.365      -0.745       0.319\n",
      "likes              1.4794      0.794      1.862      0.112      -0.464       3.423\n",
      "retweet           -0.4985      0.531     -0.938      0.384      -1.798       0.801\n",
      "count_comments     1.2680      0.250      5.069      0.002       0.656       1.880\n",
      "senti_negative    -0.0727      0.511     -0.142      0.892      -1.324       1.178\n",
      "senti_neutral     -0.3358      0.671     -0.501      0.634      -1.977       1.305\n",
      "senti_positive    -1.3056      0.592     -2.207      0.069      -2.753       0.142\n",
      "==============================================================================\n",
      "Omnibus:                        4.119   Durbin-Watson:                   1.343\n",
      "Prob(Omnibus):                  0.128   Jarque-Bera (JB):                1.665\n",
      "Skew:                          -0.820   Prob(JB):                        0.435\n",
      "Kurtosis:                       3.620   Cond. No.                         13.9\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\scipy\\stats\\stats.py:1394: UserWarning: kurtosistest only valid for n>=20 ... continuing anyway, n=13\n",
      "  \"anyway, n=%i\" % int(n))\n"
     ]
    }
   ],
   "source": [
    "# Create the linear model and complete the least squares fit\n",
    "model = sm.OLS(train_targets, train_features)\n",
    "results = model.fit()  # fit the model\n",
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "const             0.365398\n",
      "likes             0.111865\n",
      "retweet           0.384302\n",
      "count_comments    0.002289\n",
      "senti_negative    0.891617\n",
      "senti_neutral     0.634358\n",
      "senti_positive    0.069440\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# Examine pvalues\n",
    "# Features with p <= 0.05 are typically considered significantly different from 0\n",
    "print(results.pvalues)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make predictions from our model for train and test sets\n",
    "train_predictions = results.predict(train_features)\n",
    "test_predictions = results.predict(test_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Scatter the predictions vs the targets with 80% transparency\n",
    "plt.scatter(train_predictions, train_targets, alpha=0.2, color='b', label='train')\n",
    "plt.scatter(test_predictions, test_targets, alpha=0.2, color='r', label='test')\n",
    "\n",
    "# Plot the perfect prediction line\n",
    "xmin, xmax = plt.xlim()\n",
    "plt.plot(np.arange(xmin, xmax, 0.01), np.arange(xmin, xmax, 0.01), c='k')\n",
    "\n",
    "# Set the axis labels and show the plot\n",
    "plt.xlabel('predictions')\n",
    "plt.ylabel('actual')\n",
    "plt.legend()  # show the legend\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8384643116951802\n",
      "-1.6848653996684244\n"
     ]
    }
   ],
   "source": [
    "# Try with Random Forest\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "# Create the random forest model and fit to the training data\n",
    "rfr = RandomForestRegressor(n_estimators=200)\n",
    "rfr.fit(train_features, train_targets)\n",
    "\n",
    "# Look at the R^2 scores on train and test\n",
    "print(rfr.score(train_features, train_targets))\n",
    "print(rfr.score(test_features, test_targets))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Get feature importances from our random forest model\n",
    "importances = rfr.feature_importances_\n",
    "\n",
    "#importances = importances[1:]\n",
    "\n",
    "# Get the index of importances from greatest importance to least\n",
    "sorted_index = np.argsort(importances)[::-1]\n",
    "x = range(len(importances))\n",
    "\n",
    "# Create tick labels \n",
    "labels = np.array(tweet_fund_df_nom.columns.values)[sorted_index]\n",
    "plt.bar(x, importances[sorted_index], tick_label=labels)\n",
    "\n",
    "# Rotate tick labels to vertical\n",
    "plt.xticks(rotation=90)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8070233126623585\n",
      "-1.6743921181243215\n"
     ]
    }
   ],
   "source": [
    "#try out gradient boosting model\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "\n",
    "# Create GB model -- hyperparameters have already been searched for you\n",
    "gbr = GradientBoostingRegressor(max_features=4,\n",
    "                                learning_rate=0.01,\n",
    "                                n_estimators=200,\n",
    "                                subsample=0.6,\n",
    "                                random_state=42)\n",
    "gbr.fit(train_features, train_targets)\n",
    "\n",
    "\n",
    "print(gbr.score(train_features, train_targets))\n",
    "print(gbr.score(test_features, test_targets))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE: 1.174823\n",
      "0.999981440932979\n",
      "-2.7679015038715566\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\xgboost\\core.py:587: FutureWarning: Series.base is deprecated and will be removed in a future version\n",
      "  if getattr(data, 'base', None) is not None and \\\n",
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\xgboost\\core.py:588: FutureWarning: Series.base is deprecated and will be removed in a future version\n",
      "  data.base is not None and isinstance(data, np.ndarray) \\\n"
     ]
    }
   ],
   "source": [
    "import xgboost as xgb\n",
    "from sklearn.metrics import mean_squared_error\n",
    "# Instantiate the XGBRegressor: xg_reg\n",
    "xg_reg = xgb.XGBRegressor(seed=123, objective='reg:linear', max_depth=10)\n",
    "\n",
    "# Fit the regressor to the training set\n",
    "xg_reg.fit(train_features, train_targets)\n",
    "\n",
    "# Predict the labels of the test set: preds\n",
    "preds = xg_reg.predict(test_features)\n",
    "\n",
    "# Compute the rmse: rmse\n",
    "rmse = np.sqrt(mean_squared_error(test_targets, preds))\n",
    "print(\"RMSE: %f\" % (rmse))\n",
    "\n",
    "\n",
    "print(xg_reg.score(train_features, train_targets))\n",
    "print(xg_reg.score(test_features, test_targets))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### WebTraffic Analysis with Revenue data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load web traffic data extract from altcap database\n",
    "web_traf = pd.read_csv('C:\\\\Users\\\\eche1\\\\Downloads\\\\BKNG_webtraffic.csv', parse_dates = ['date'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "        date  traffic  paid_traffic\n",
      "0 2012-01-15   339352        429520\n",
      "1 2012-02-15   305865        492690\n",
      "2 2012-03-15   261274        481117\n",
      "3 2012-04-15   354461        472161\n",
      "4 2012-05-15   348095        344127\n",
      "\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 87 entries, 0 to 86\n",
      "Data columns (total 3 columns):\n",
      "date            87 non-null datetime64[ns]\n",
      "traffic         87 non-null int64\n",
      "paid_traffic    87 non-null int64\n",
      "dtypes: datetime64[ns](1), int64(2)\n",
      "memory usage: 2.1 KB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "#Explore data\n",
    "print(web_traf.head())\n",
    "print()\n",
    "print(web_traf.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "#set date as index for later use in dataframe time series, resampling (quarterly) and  join operations\n",
    "web_traf.index = pd.to_datetime(web_traf.date)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Resample the data fram to Quarterly to match the revenue data time series\n",
    "web_traf_quarterly = web_traf.resample('Q', label='right').sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "            traffic  paid_traffic\n",
      "date                             \n",
      "2012-03-31   906491       1403327\n",
      "2012-06-30   938977       1203557\n",
      "2012-09-30  1049085        603426\n",
      "2012-12-31  1374946       1213900\n",
      "2013-03-31  1795384       1564678\n",
      "\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "DatetimeIndex: 29 entries, 2012-03-31 to 2019-03-31\n",
      "Freq: Q-DEC\n",
      "Data columns (total 2 columns):\n",
      "traffic         29 non-null int64\n",
      "paid_traffic    29 non-null int64\n",
      "dtypes: int64(2)\n",
      "memory usage: 696.0 bytes\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "#Explore data\n",
    "print(web_traf_quarterly.head())\n",
    "print()\n",
    "print(web_traf_quarterly.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Join web traffic and revenue dataframe by quarterly date\n",
    "#webtraffic_fund_df =  fund_df[['totalRevenue','costOfRevenue']].join(web_traf_quarterly,how='inner')\n",
    "webtraffic_fund_df = fund_df[['totalRevenue']].join(web_traf_quarterly,how='inner')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             totalRevenue   traffic  paid_traffic\n",
      "date                                             \n",
      "2019-03-31  2837000000.00  41952583     292987022\n",
      "2018-12-31  3212615000.00  43465557     197796997\n",
      "2018-09-30  4849090000.00  39868199      85293585\n",
      "2018-06-30  3537094000.00  43277687      16451079\n",
      "2018-03-31  2928201000.00  42974170      15235298\n",
      "\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "DatetimeIndex: 29 entries, 2019-03-31 to 2012-03-31\n",
      "Freq: -1Q-DEC\n",
      "Data columns (total 3 columns):\n",
      "totalRevenue    29 non-null object\n",
      "traffic         29 non-null int64\n",
      "paid_traffic    29 non-null int64\n",
      "dtypes: int64(2), object(1)\n",
      "memory usage: 928.0+ bytes\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "#Explore data\n",
    "print(webtraffic_fund_df.head())\n",
    "print()\n",
    "print(webtraffic_fund_df.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:625: DataConversionWarning: Data with input dtype int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\sklearn\\base.py:462: DataConversionWarning: Data with input dtype int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    }
   ],
   "source": [
    "#Standardise and normalise datafram values\n",
    "from sklearn import preprocessing as prep\n",
    "#tweet_fund_df_nom = pd.DataFrame(prep.MinMaxScaler(feature_range=(0, 1)).fit_transform(tweet_fund_df), columns=tweet_fund_df.columns, index=tweet_fund_df.index)\n",
    "webtraffic_fund_df_nom = pd.DataFrame(prep.StandardScaler().fit_transform(webtraffic_fund_df), columns=webtraffic_fund_df.columns, index=webtraffic_fund_df.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>totalRevenue</th>\n",
       "      <th>traffic</th>\n",
       "      <th>paid_traffic</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-03-31</th>\n",
       "      <td>0.442211</td>\n",
       "      <td>1.860641</td>\n",
       "      <td>4.240993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-12-31</th>\n",
       "      <td>0.851651</td>\n",
       "      <td>1.963628</td>\n",
       "      <td>2.736939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-30</th>\n",
       "      <td>2.635494</td>\n",
       "      <td>1.718759</td>\n",
       "      <td>0.959324</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-30</th>\n",
       "      <td>1.205350</td>\n",
       "      <td>1.950840</td>\n",
       "      <td>-0.128425</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-03-31</th>\n",
       "      <td>0.541625</td>\n",
       "      <td>1.930180</td>\n",
       "      <td>-0.147635</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            totalRevenue   traffic  paid_traffic\n",
       "date                                            \n",
       "2019-03-31      0.442211  1.860641      4.240993\n",
       "2018-12-31      0.851651  1.963628      2.736939\n",
       "2018-09-30      2.635494  1.718759      0.959324\n",
       "2018-06-30      1.205350  1.950840     -0.128425\n",
       "2018-03-31      0.541625  1.930180     -0.147635"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore standardise dataframe\n",
    "webtraffic_fund_df_nom.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "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>totalRevenue</th>\n",
       "      <th>traffic</th>\n",
       "      <th>paid_traffic</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>totalRevenue</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.725199</td>\n",
       "      <td>0.321129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>traffic</th>\n",
       "      <td>0.725199</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.626429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>paid_traffic</th>\n",
       "      <td>0.321129</td>\n",
       "      <td>0.626429</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              totalRevenue   traffic  paid_traffic\n",
       "totalRevenue      1.000000  0.725199      0.321129\n",
       "traffic           0.725199  1.000000      0.626429\n",
       "paid_traffic      0.321129  0.626429      1.000000"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore correlation between Revenue and other feature values\n",
    "webtraffic_fund_df_nom.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([<matplotlib.axes._subplots.AxesSubplot object at 0x0000017B96241518>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B96261DA0>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B9628D5F8>],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Explore visual correlation between Revenue and other feature values\n",
    "webtraffic_fund_df_nom.plot(subplots=True,figsize=(10,6))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'traffic_0': 0.7251990769072048},\n",
       " {'traffic_1': 0.7578193971300717},\n",
       " {'traffic_2': 0.8070296948415889},\n",
       " {'traffic_3': 0.7833742197780313},\n",
       " {'traffic_4': 0.710161961990801},\n",
       " {'paid_traffic_0': 0.32112942442565384},\n",
       " {'paid_traffic_1': 0.4997085398623814},\n",
       " {'paid_traffic_2': 0.6246249703818663},\n",
       " {'paid_traffic_3': 0.4193649061391939},\n",
       " {'paid_traffic_4': 0.3963421842676794}]"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Any correlation by shifting or lagging the time series data?\n",
    "feats = webtraffic_fund_df_nom.drop('totalRevenue', axis=1).columns.values\n",
    "[ {x + '_' + str(y) : webtraffic_fund_df_nom[x].shift(y).corr(webtraffic_fund_df_nom.totalRevenue)} for x in feats for y in range(5) ]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Prepare data for Stats and Machine Learning models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((29, 2), (29,))"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Select target(dependent/label) and feature(independent) variables\n",
    "targets = webtraffic_fund_df_nom['totalRevenue']\n",
    "features = webtraffic_fund_df_nom.drop(['totalRevenue'], axis=1)\n",
    "feature_names = features.columns\n",
    "features.shape, targets.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['const', 'traffic', 'paid_traffic'], dtype='object')"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Using a linear OLS model here \n",
    "import statsmodels.api as sm\n",
    "# Add a constant to the features\n",
    "linear_features = sm.add_constant(features)\n",
    "linear_features.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(29, 3) (24, 3) (5, 3)\n"
     ]
    }
   ],
   "source": [
    "# Create a size for the training set that is 85% of the total number of samples\n",
    "train_size = int(0.85 * targets.shape[0])\n",
    "train_features = linear_features[:train_size]\n",
    "train_targets = targets[:train_size]\n",
    "test_features = linear_features[train_size:]\n",
    "test_targets = targets[train_size:]\n",
    "print(linear_features.shape, train_features.shape, test_features.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:           totalRevenue   R-squared:                       0.460\n",
      "Model:                            OLS   Adj. R-squared:                  0.409\n",
      "Method:                 Least Squares   F-statistic:                     8.947\n",
      "Date:                Mon, 17 Jun 2019   Prob (F-statistic):            0.00155\n",
      "Time:                        00:40:06   Log-Likelihood:                -24.304\n",
      "No. Observations:                  24   AIC:                             54.61\n",
      "Df Residuals:                      21   BIC:                             58.14\n",
      "Df Model:                           2                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "================================================================================\n",
      "                   coef    std err          t      P>|t|      [0.025      0.975]\n",
      "--------------------------------------------------------------------------------\n",
      "const            0.1316      0.148      0.888      0.385      -0.177       0.440\n",
      "traffic          0.7149      0.186      3.853      0.001       0.329       1.101\n",
      "paid_traffic    -0.1751      0.171     -1.024      0.318      -0.531       0.181\n",
      "==============================================================================\n",
      "Omnibus:                        6.092   Durbin-Watson:                   2.149\n",
      "Prob(Omnibus):                  0.048   Jarque-Bera (JB):                4.764\n",
      "Skew:                           1.088   Prob(JB):                       0.0924\n",
      "Kurtosis:                       3.163   Cond. No.                         2.14\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    }
   ],
   "source": [
    "# Create the linear model and complete the least squares fit\n",
    "model = sm.OLS(train_targets, train_features)\n",
    "results = model.fit()  # fit the model\n",
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "const           0.384545\n",
      "traffic         0.000922\n",
      "paid_traffic    0.317671\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# examine pvalues\n",
    "# Features with p <= 0.05 are typically considered significantly different from 0\n",
    "print(results.pvalues)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make predictions from our model for train and test sets\n",
    "train_predictions = results.predict(train_features)\n",
    "test_predictions = results.predict(test_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Scatter the predictions vs the targets with 80% transparency\n",
    "plt.scatter(train_predictions, train_targets, alpha=0.2, color='b', label='train')\n",
    "plt.scatter(test_predictions, test_targets, alpha=0.2, color='r', label='test')\n",
    "\n",
    "# Plot the perfect prediction line\n",
    "xmin, xmax = plt.xlim()\n",
    "plt.plot(np.arange(xmin, xmax, 0.01), np.arange(xmin, xmax, 0.01), c='k')\n",
    "\n",
    "# Set the axis labels and show the plot\n",
    "plt.xlabel('predictions')\n",
    "plt.ylabel('actual')\n",
    "plt.legend()  # show the legend\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.0\n",
      "-2.747050776207829\n"
     ]
    }
   ],
   "source": [
    "## trying Decision Tree Models\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "\n",
    "# Create a decision tree regression model with default arguments\n",
    "decision_tree = DecisionTreeRegressor()\n",
    "\n",
    "# Fit the model to the training features and targets\n",
    "decision_tree.fit(train_features, train_targets)\n",
    "\n",
    "# Check the score on train and test\n",
    "print(decision_tree.score(train_features, train_targets))\n",
    "print(decision_tree.score(test_features, test_targets))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9111840523724524\n",
      "-4.748078787147845\n"
     ]
    }
   ],
   "source": [
    "# Try with Random Forest\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "# Create the random forest model and fit to the training data\n",
    "rfr = RandomForestRegressor(n_estimators=200)\n",
    "rfr.fit(train_features, train_targets)\n",
    "\n",
    "# Look at the R^2 scores on train and test\n",
    "print(rfr.score(train_features, train_targets))\n",
    "print(rfr.score(test_features, test_targets))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Get feature importances from our random forest model\n",
    "importances = rfr.feature_importances_\n",
    "\n",
    "importances = importances[1:]\n",
    "\n",
    "# Get the index of importances from greatest importance to least\n",
    "sorted_index = np.argsort(importances)[::-1]\n",
    "x = range(len(importances))\n",
    "\n",
    "# Create tick labels \n",
    "labels = np.array(feature_names)[sorted_index]\n",
    "plt.bar(x, importances[sorted_index], tick_label=labels)\n",
    "\n",
    "# Rotate tick labels to vertical\n",
    "plt.xticks(rotation=90)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8847358086224316\n",
      "-7.679102696026689\n"
     ]
    }
   ],
   "source": [
    "#try out gradient boosting model\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "\n",
    "# Create GB model -- hyperparameters have already been searched for you\n",
    "gbr = GradientBoostingRegressor(learning_rate=0.01,\n",
    "                                n_estimators=200,\n",
    "                                subsample=0.6,\n",
    "                                random_state=42)\n",
    "gbr.fit(train_features, train_targets)\n",
    "\n",
    "print(gbr.score(train_features, train_targets))\n",
    "print(gbr.score(test_features, test_targets))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Combine alternative data from Webtraffic and Twitter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "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>totalRevenue</th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-03-31</th>\n",
       "      <td>2837000000.00</td>\n",
       "      <td>302</td>\n",
       "      <td>53</td>\n",
       "      <td>285</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-12-31</th>\n",
       "      <td>3212615000.00</td>\n",
       "      <td>246</td>\n",
       "      <td>66</td>\n",
       "      <td>260</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-30</th>\n",
       "      <td>4849090000.00</td>\n",
       "      <td>344</td>\n",
       "      <td>68</td>\n",
       "      <td>385</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-30</th>\n",
       "      <td>3537094000.00</td>\n",
       "      <td>897</td>\n",
       "      <td>112</td>\n",
       "      <td>217</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-03-31</th>\n",
       "      <td>2928201000.00</td>\n",
       "      <td>1199</td>\n",
       "      <td>159</td>\n",
       "      <td>222</td>\n",
       "      <td>9</td>\n",
       "      <td>27</td>\n",
       "      <td>62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31</th>\n",
       "      <td>2803093000.00</td>\n",
       "      <td>1492</td>\n",
       "      <td>303</td>\n",
       "      <td>500</td>\n",
       "      <td>8</td>\n",
       "      <td>34</td>\n",
       "      <td>105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-09-30</th>\n",
       "      <td>4434029000.00</td>\n",
       "      <td>1160</td>\n",
       "      <td>250</td>\n",
       "      <td>679</td>\n",
       "      <td>9</td>\n",
       "      <td>32</td>\n",
       "      <td>82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-06-30</th>\n",
       "      <td>3024556000.00</td>\n",
       "      <td>742</td>\n",
       "      <td>183</td>\n",
       "      <td>313</td>\n",
       "      <td>2</td>\n",
       "      <td>13</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-31</th>\n",
       "      <td>2419404000.00</td>\n",
       "      <td>494</td>\n",
       "      <td>130</td>\n",
       "      <td>64</td>\n",
       "      <td>4</td>\n",
       "      <td>7</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31</th>\n",
       "      <td>2348433000.00</td>\n",
       "      <td>556</td>\n",
       "      <td>120</td>\n",
       "      <td>45</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-09-30</th>\n",
       "      <td>3690552000.00</td>\n",
       "      <td>1051</td>\n",
       "      <td>327</td>\n",
       "      <td>278</td>\n",
       "      <td>3</td>\n",
       "      <td>16</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-06-30</th>\n",
       "      <td>2555902000.00</td>\n",
       "      <td>585</td>\n",
       "      <td>141</td>\n",
       "      <td>64</td>\n",
       "      <td>3</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-03-31</th>\n",
       "      <td>2148119000.00</td>\n",
       "      <td>171</td>\n",
       "      <td>85</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-12-31</th>\n",
       "      <td>1999995000.00</td>\n",
       "      <td>154</td>\n",
       "      <td>57</td>\n",
       "      <td>35</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-09-30</th>\n",
       "      <td>3102901000.00</td>\n",
       "      <td>263</td>\n",
       "      <td>190</td>\n",
       "      <td>139</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-06-30</th>\n",
       "      <td>2280397000.00</td>\n",
       "      <td>896</td>\n",
       "      <td>348</td>\n",
       "      <td>152</td>\n",
       "      <td>4</td>\n",
       "      <td>39</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             totalRevenue  likes  retweet  count_comments  senti_negative  \\\n",
       "2019-03-31  2837000000.00    302       53             285               2   \n",
       "2018-12-31  3212615000.00    246       66             260               0   \n",
       "2018-09-30  4849090000.00    344       68             385               0   \n",
       "2018-06-30  3537094000.00    897      112             217               5   \n",
       "2018-03-31  2928201000.00   1199      159             222               9   \n",
       "2017-12-31  2803093000.00   1492      303             500               8   \n",
       "2017-09-30  4434029000.00   1160      250             679               9   \n",
       "2017-06-30  3024556000.00    742      183             313               2   \n",
       "2017-03-31  2419404000.00    494      130              64               4   \n",
       "2016-12-31  2348433000.00    556      120              45               2   \n",
       "2016-09-30  3690552000.00   1051      327             278               3   \n",
       "2016-06-30  2555902000.00    585      141              64               3   \n",
       "2016-03-31  2148119000.00    171       85              15               0   \n",
       "2015-12-31  1999995000.00    154       57              35               1   \n",
       "2015-09-30  3102901000.00    263      190             139               0   \n",
       "2015-06-30  2280397000.00    896      348             152               4   \n",
       "\n",
       "            senti_neutral  senti_positive  \n",
       "2019-03-31              0              11  \n",
       "2018-12-31              2               8  \n",
       "2018-09-30              2               8  \n",
       "2018-06-30              5              50  \n",
       "2018-03-31             27              62  \n",
       "2017-12-31             34             105  \n",
       "2017-09-30             32              82  \n",
       "2017-06-30             13              47  \n",
       "2017-03-31              7              20  \n",
       "2016-12-31              7              14  \n",
       "2016-09-30             16              27  \n",
       "2016-06-30              6              12  \n",
       "2016-03-31              3               3  \n",
       "2015-12-31              9               9  \n",
       "2015-09-30             10              17  \n",
       "2015-06-30             39              48  "
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Review tweet and revenue data frame\n",
    "tweet_fund_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "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>totalRevenue</th>\n",
       "      <th>traffic</th>\n",
       "      <th>paid_traffic</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>date</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-03-31</th>\n",
       "      <td>2837000000.00</td>\n",
       "      <td>41952583</td>\n",
       "      <td>292987022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-12-31</th>\n",
       "      <td>3212615000.00</td>\n",
       "      <td>43465557</td>\n",
       "      <td>197796997</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-30</th>\n",
       "      <td>4849090000.00</td>\n",
       "      <td>39868199</td>\n",
       "      <td>85293585</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-30</th>\n",
       "      <td>3537094000.00</td>\n",
       "      <td>43277687</td>\n",
       "      <td>16451079</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-03-31</th>\n",
       "      <td>2928201000.00</td>\n",
       "      <td>42974170</td>\n",
       "      <td>15235298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31</th>\n",
       "      <td>2803093000.00</td>\n",
       "      <td>29406354</td>\n",
       "      <td>9355158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-09-30</th>\n",
       "      <td>4434029000.00</td>\n",
       "      <td>25912258</td>\n",
       "      <td>18947528</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-06-30</th>\n",
       "      <td>3024556000.00</td>\n",
       "      <td>24841279</td>\n",
       "      <td>16137900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-31</th>\n",
       "      <td>2419404000.00</td>\n",
       "      <td>20686396</td>\n",
       "      <td>7537993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31</th>\n",
       "      <td>2348433000.00</td>\n",
       "      <td>15091975</td>\n",
       "      <td>1480949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-09-30</th>\n",
       "      <td>3690552000.00</td>\n",
       "      <td>10727931</td>\n",
       "      <td>2939493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-06-30</th>\n",
       "      <td>2555902000.00</td>\n",
       "      <td>9685703</td>\n",
       "      <td>2716816</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-03-31</th>\n",
       "      <td>2148119000.00</td>\n",
       "      <td>8749933</td>\n",
       "      <td>3583808</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-12-31</th>\n",
       "      <td>1999995000.00</td>\n",
       "      <td>8968709</td>\n",
       "      <td>2789117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-09-30</th>\n",
       "      <td>3102901000.00</td>\n",
       "      <td>8351469</td>\n",
       "      <td>3099355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-06-30</th>\n",
       "      <td>2280397000.00</td>\n",
       "      <td>6907676</td>\n",
       "      <td>4255740</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-31</th>\n",
       "      <td>1840694000.00</td>\n",
       "      <td>6627443</td>\n",
       "      <td>2105389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2014-12-31</th>\n",
       "      <td>1840097000.00</td>\n",
       "      <td>6205281</td>\n",
       "      <td>4280217</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2014-09-30</th>\n",
       "      <td>2836497000.00</td>\n",
       "      <td>6424505</td>\n",
       "      <td>3892952</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2014-06-30</th>\n",
       "      <td>2123575000.00</td>\n",
       "      <td>5389961</td>\n",
       "      <td>4652138</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2014-03-31</th>\n",
       "      <td>1641802000.00</td>\n",
       "      <td>4871828</td>\n",
       "      <td>4629767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013-12-31</th>\n",
       "      <td>1541153000.00</td>\n",
       "      <td>3618230</td>\n",
       "      <td>3483105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013-09-30</th>\n",
       "      <td>2269903000.00</td>\n",
       "      <td>2123082</td>\n",
       "      <td>1738719</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013-06-30</th>\n",
       "      <td>1680238000.00</td>\n",
       "      <td>1730913</td>\n",
       "      <td>1411068</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2013-03-31</th>\n",
       "      <td>1302012000.00</td>\n",
       "      <td>1795384</td>\n",
       "      <td>1564678</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-12-31</th>\n",
       "      <td>1190640000.00</td>\n",
       "      <td>1374946</td>\n",
       "      <td>1213900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-09-30</th>\n",
       "      <td>1706310000.00</td>\n",
       "      <td>1049085</td>\n",
       "      <td>603426</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-06-30</th>\n",
       "      <td>1326759000.00</td>\n",
       "      <td>938977</td>\n",
       "      <td>1203557</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-03-31</th>\n",
       "      <td>1037247000.00</td>\n",
       "      <td>906491</td>\n",
       "      <td>1403327</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             totalRevenue   traffic  paid_traffic\n",
       "date                                             \n",
       "2019-03-31  2837000000.00  41952583     292987022\n",
       "2018-12-31  3212615000.00  43465557     197796997\n",
       "2018-09-30  4849090000.00  39868199      85293585\n",
       "2018-06-30  3537094000.00  43277687      16451079\n",
       "2018-03-31  2928201000.00  42974170      15235298\n",
       "2017-12-31  2803093000.00  29406354       9355158\n",
       "2017-09-30  4434029000.00  25912258      18947528\n",
       "2017-06-30  3024556000.00  24841279      16137900\n",
       "2017-03-31  2419404000.00  20686396       7537993\n",
       "2016-12-31  2348433000.00  15091975       1480949\n",
       "2016-09-30  3690552000.00  10727931       2939493\n",
       "2016-06-30  2555902000.00   9685703       2716816\n",
       "2016-03-31  2148119000.00   8749933       3583808\n",
       "2015-12-31  1999995000.00   8968709       2789117\n",
       "2015-09-30  3102901000.00   8351469       3099355\n",
       "2015-06-30  2280397000.00   6907676       4255740\n",
       "2015-03-31  1840694000.00   6627443       2105389\n",
       "2014-12-31  1840097000.00   6205281       4280217\n",
       "2014-09-30  2836497000.00   6424505       3892952\n",
       "2014-06-30  2123575000.00   5389961       4652138\n",
       "2014-03-31  1641802000.00   4871828       4629767\n",
       "2013-12-31  1541153000.00   3618230       3483105\n",
       "2013-09-30  2269903000.00   2123082       1738719\n",
       "2013-06-30  1680238000.00   1730913       1411068\n",
       "2013-03-31  1302012000.00   1795384       1564678\n",
       "2012-12-31  1190640000.00   1374946       1213900\n",
       "2012-09-30  1706310000.00   1049085        603426\n",
       "2012-06-30  1326759000.00    938977       1203557\n",
       "2012-03-31  1037247000.00    906491       1403327"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Review web traffic and revenue data frame\n",
    "webtraffic_fund_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "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>traffic</th>\n",
       "      <th>paid_traffic</th>\n",
       "      <th>totalRevenue</th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-03-31</th>\n",
       "      <td>41952583</td>\n",
       "      <td>292987022</td>\n",
       "      <td>2837000000.00</td>\n",
       "      <td>302</td>\n",
       "      <td>53</td>\n",
       "      <td>285</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-12-31</th>\n",
       "      <td>43465557</td>\n",
       "      <td>197796997</td>\n",
       "      <td>3212615000.00</td>\n",
       "      <td>246</td>\n",
       "      <td>66</td>\n",
       "      <td>260</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-30</th>\n",
       "      <td>39868199</td>\n",
       "      <td>85293585</td>\n",
       "      <td>4849090000.00</td>\n",
       "      <td>344</td>\n",
       "      <td>68</td>\n",
       "      <td>385</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-30</th>\n",
       "      <td>43277687</td>\n",
       "      <td>16451079</td>\n",
       "      <td>3537094000.00</td>\n",
       "      <td>897</td>\n",
       "      <td>112</td>\n",
       "      <td>217</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-03-31</th>\n",
       "      <td>42974170</td>\n",
       "      <td>15235298</td>\n",
       "      <td>2928201000.00</td>\n",
       "      <td>1199</td>\n",
       "      <td>159</td>\n",
       "      <td>222</td>\n",
       "      <td>9</td>\n",
       "      <td>27</td>\n",
       "      <td>62</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31</th>\n",
       "      <td>29406354</td>\n",
       "      <td>9355158</td>\n",
       "      <td>2803093000.00</td>\n",
       "      <td>1492</td>\n",
       "      <td>303</td>\n",
       "      <td>500</td>\n",
       "      <td>8</td>\n",
       "      <td>34</td>\n",
       "      <td>105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-09-30</th>\n",
       "      <td>25912258</td>\n",
       "      <td>18947528</td>\n",
       "      <td>4434029000.00</td>\n",
       "      <td>1160</td>\n",
       "      <td>250</td>\n",
       "      <td>679</td>\n",
       "      <td>9</td>\n",
       "      <td>32</td>\n",
       "      <td>82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-06-30</th>\n",
       "      <td>24841279</td>\n",
       "      <td>16137900</td>\n",
       "      <td>3024556000.00</td>\n",
       "      <td>742</td>\n",
       "      <td>183</td>\n",
       "      <td>313</td>\n",
       "      <td>2</td>\n",
       "      <td>13</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-31</th>\n",
       "      <td>20686396</td>\n",
       "      <td>7537993</td>\n",
       "      <td>2419404000.00</td>\n",
       "      <td>494</td>\n",
       "      <td>130</td>\n",
       "      <td>64</td>\n",
       "      <td>4</td>\n",
       "      <td>7</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31</th>\n",
       "      <td>15091975</td>\n",
       "      <td>1480949</td>\n",
       "      <td>2348433000.00</td>\n",
       "      <td>556</td>\n",
       "      <td>120</td>\n",
       "      <td>45</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-09-30</th>\n",
       "      <td>10727931</td>\n",
       "      <td>2939493</td>\n",
       "      <td>3690552000.00</td>\n",
       "      <td>1051</td>\n",
       "      <td>327</td>\n",
       "      <td>278</td>\n",
       "      <td>3</td>\n",
       "      <td>16</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-06-30</th>\n",
       "      <td>9685703</td>\n",
       "      <td>2716816</td>\n",
       "      <td>2555902000.00</td>\n",
       "      <td>585</td>\n",
       "      <td>141</td>\n",
       "      <td>64</td>\n",
       "      <td>3</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-03-31</th>\n",
       "      <td>8749933</td>\n",
       "      <td>3583808</td>\n",
       "      <td>2148119000.00</td>\n",
       "      <td>171</td>\n",
       "      <td>85</td>\n",
       "      <td>15</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-12-31</th>\n",
       "      <td>8968709</td>\n",
       "      <td>2789117</td>\n",
       "      <td>1999995000.00</td>\n",
       "      <td>154</td>\n",
       "      <td>57</td>\n",
       "      <td>35</td>\n",
       "      <td>1</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-09-30</th>\n",
       "      <td>8351469</td>\n",
       "      <td>3099355</td>\n",
       "      <td>3102901000.00</td>\n",
       "      <td>263</td>\n",
       "      <td>190</td>\n",
       "      <td>139</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-06-30</th>\n",
       "      <td>6907676</td>\n",
       "      <td>4255740</td>\n",
       "      <td>2280397000.00</td>\n",
       "      <td>896</td>\n",
       "      <td>348</td>\n",
       "      <td>152</td>\n",
       "      <td>4</td>\n",
       "      <td>39</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             traffic  paid_traffic   totalRevenue  likes  retweet  \\\n",
       "2019-03-31  41952583     292987022  2837000000.00    302       53   \n",
       "2018-12-31  43465557     197796997  3212615000.00    246       66   \n",
       "2018-09-30  39868199      85293585  4849090000.00    344       68   \n",
       "2018-06-30  43277687      16451079  3537094000.00    897      112   \n",
       "2018-03-31  42974170      15235298  2928201000.00   1199      159   \n",
       "2017-12-31  29406354       9355158  2803093000.00   1492      303   \n",
       "2017-09-30  25912258      18947528  4434029000.00   1160      250   \n",
       "2017-06-30  24841279      16137900  3024556000.00    742      183   \n",
       "2017-03-31  20686396       7537993  2419404000.00    494      130   \n",
       "2016-12-31  15091975       1480949  2348433000.00    556      120   \n",
       "2016-09-30  10727931       2939493  3690552000.00   1051      327   \n",
       "2016-06-30   9685703       2716816  2555902000.00    585      141   \n",
       "2016-03-31   8749933       3583808  2148119000.00    171       85   \n",
       "2015-12-31   8968709       2789117  1999995000.00    154       57   \n",
       "2015-09-30   8351469       3099355  3102901000.00    263      190   \n",
       "2015-06-30   6907676       4255740  2280397000.00    896      348   \n",
       "\n",
       "            count_comments  senti_negative  senti_neutral  senti_positive  \n",
       "2019-03-31             285               2              0              11  \n",
       "2018-12-31             260               0              2               8  \n",
       "2018-09-30             385               0              2               8  \n",
       "2018-06-30             217               5              5              50  \n",
       "2018-03-31             222               9             27              62  \n",
       "2017-12-31             500               8             34             105  \n",
       "2017-09-30             679               9             32              82  \n",
       "2017-06-30             313               2             13              47  \n",
       "2017-03-31              64               4              7              20  \n",
       "2016-12-31              45               2              7              14  \n",
       "2016-09-30             278               3             16              27  \n",
       "2016-06-30              64               3              6              12  \n",
       "2016-03-31              15               0              3               3  \n",
       "2015-12-31              35               1              9               9  \n",
       "2015-09-30             139               0             10              17  \n",
       "2015-06-30             152               4             39              48  "
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore joining tweet_fund_df to webtraffic_fund_df  \n",
    "webtraffic_fund_df[['traffic','paid_traffic']].join(tweet_fund_df,how='inner')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Combine tweet_fund_df to webtraffic_fund_df \n",
    "webtraffic_tweet_fund_df =  webtraffic_fund_df[['traffic','paid_traffic']].join(tweet_fund_df,how='inner')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\sklearn\\preprocessing\\data.py:625: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.partial_fit(X, y)\n",
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\sklearn\\base.py:462: DataConversionWarning: Data with input dtype uint8, int64, object were all converted to float64 by StandardScaler.\n",
      "  return self.fit(X, **fit_params).transform(X)\n"
     ]
    }
   ],
   "source": [
    "#Standardise and normalise dataframe values\n",
    "from sklearn import preprocessing as prep\n",
    "#tweet_fund_df_nom = pd.DataFrame(prep.MinMaxScaler(feature_range=(0, 1)).fit_transform(tweet_fund_df), columns=tweet_fund_df.columns, index=tweet_fund_df.index)\n",
    "webtraffic_tweet_fund_df_nom = pd.DataFrame(prep.StandardScaler().fit_transform(webtraffic_tweet_fund_df), columns=webtraffic_tweet_fund_df.columns, index=webtraffic_tweet_fund_df.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "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>traffic</th>\n",
       "      <th>paid_traffic</th>\n",
       "      <th>totalRevenue</th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2019-03-31</th>\n",
       "      <td>1.286393</td>\n",
       "      <td>3.097198</td>\n",
       "      <td>-0.224631</td>\n",
       "      <td>-0.886782</td>\n",
       "      <td>-1.151780</td>\n",
       "      <td>0.320742</td>\n",
       "      <td>-0.415227</td>\n",
       "      <td>-1.081182</td>\n",
       "      <td>-0.743672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-12-31</th>\n",
       "      <td>1.393636</td>\n",
       "      <td>1.920023</td>\n",
       "      <td>0.261087</td>\n",
       "      <td>-1.025690</td>\n",
       "      <td>-1.014412</td>\n",
       "      <td>0.179290</td>\n",
       "      <td>-1.079591</td>\n",
       "      <td>-0.917985</td>\n",
       "      <td>-0.846543</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-09-30</th>\n",
       "      <td>1.138647</td>\n",
       "      <td>0.528741</td>\n",
       "      <td>2.377258</td>\n",
       "      <td>-0.782600</td>\n",
       "      <td>-0.993278</td>\n",
       "      <td>0.886550</td>\n",
       "      <td>-1.079591</td>\n",
       "      <td>-0.917985</td>\n",
       "      <td>-0.846543</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-06-30</th>\n",
       "      <td>1.380319</td>\n",
       "      <td>-0.322606</td>\n",
       "      <td>0.680680</td>\n",
       "      <td>0.589121</td>\n",
       "      <td>-0.528339</td>\n",
       "      <td>-0.064007</td>\n",
       "      <td>0.581318</td>\n",
       "      <td>-0.673189</td>\n",
       "      <td>0.593652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-03-31</th>\n",
       "      <td>1.358805</td>\n",
       "      <td>-0.337641</td>\n",
       "      <td>-0.106696</td>\n",
       "      <td>1.338234</td>\n",
       "      <td>-0.031700</td>\n",
       "      <td>-0.035717</td>\n",
       "      <td>1.910046</td>\n",
       "      <td>1.121982</td>\n",
       "      <td>1.005136</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31</th>\n",
       "      <td>0.397090</td>\n",
       "      <td>-0.410358</td>\n",
       "      <td>-0.268477</td>\n",
       "      <td>2.065023</td>\n",
       "      <td>1.489917</td>\n",
       "      <td>1.537230</td>\n",
       "      <td>1.577864</td>\n",
       "      <td>1.693173</td>\n",
       "      <td>2.479620</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-09-30</th>\n",
       "      <td>0.149421</td>\n",
       "      <td>-0.291733</td>\n",
       "      <td>1.840531</td>\n",
       "      <td>1.241495</td>\n",
       "      <td>0.929877</td>\n",
       "      <td>2.550026</td>\n",
       "      <td>1.910046</td>\n",
       "      <td>1.529975</td>\n",
       "      <td>1.690942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-06-30</th>\n",
       "      <td>0.073507</td>\n",
       "      <td>-0.326479</td>\n",
       "      <td>0.017903</td>\n",
       "      <td>0.204642</td>\n",
       "      <td>0.221903</td>\n",
       "      <td>0.479169</td>\n",
       "      <td>-0.415227</td>\n",
       "      <td>-0.020400</td>\n",
       "      <td>0.490781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-31</th>\n",
       "      <td>-0.221000</td>\n",
       "      <td>-0.432830</td>\n",
       "      <td>-0.764636</td>\n",
       "      <td>-0.410524</td>\n",
       "      <td>-0.338137</td>\n",
       "      <td>-0.929693</td>\n",
       "      <td>0.249136</td>\n",
       "      <td>-0.509992</td>\n",
       "      <td>-0.435059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-12-31</th>\n",
       "      <td>-0.617544</td>\n",
       "      <td>-0.507735</td>\n",
       "      <td>-0.856410</td>\n",
       "      <td>-0.256733</td>\n",
       "      <td>-0.443805</td>\n",
       "      <td>-1.037197</td>\n",
       "      <td>-0.415227</td>\n",
       "      <td>-0.509992</td>\n",
       "      <td>-0.640801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-09-30</th>\n",
       "      <td>-0.926877</td>\n",
       "      <td>-0.489698</td>\n",
       "      <td>0.879121</td>\n",
       "      <td>0.971119</td>\n",
       "      <td>1.743520</td>\n",
       "      <td>0.281136</td>\n",
       "      <td>-0.083045</td>\n",
       "      <td>0.224396</td>\n",
       "      <td>-0.195026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-06-30</th>\n",
       "      <td>-1.000752</td>\n",
       "      <td>-0.492452</td>\n",
       "      <td>-0.588126</td>\n",
       "      <td>-0.184798</td>\n",
       "      <td>-0.221903</td>\n",
       "      <td>-0.929693</td>\n",
       "      <td>-0.083045</td>\n",
       "      <td>-0.591590</td>\n",
       "      <td>-0.709381</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2016-03-31</th>\n",
       "      <td>-1.067081</td>\n",
       "      <td>-0.481730</td>\n",
       "      <td>-1.115442</td>\n",
       "      <td>-1.211728</td>\n",
       "      <td>-0.813643</td>\n",
       "      <td>-1.206939</td>\n",
       "      <td>-1.079591</td>\n",
       "      <td>-0.836386</td>\n",
       "      <td>-1.017995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-12-31</th>\n",
       "      <td>-1.051574</td>\n",
       "      <td>-0.491558</td>\n",
       "      <td>-1.306985</td>\n",
       "      <td>-1.253897</td>\n",
       "      <td>-1.109513</td>\n",
       "      <td>-1.093778</td>\n",
       "      <td>-0.747409</td>\n",
       "      <td>-0.346794</td>\n",
       "      <td>-0.812252</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-09-30</th>\n",
       "      <td>-1.095325</td>\n",
       "      <td>-0.487721</td>\n",
       "      <td>0.119213</td>\n",
       "      <td>-0.983522</td>\n",
       "      <td>0.295870</td>\n",
       "      <td>-0.505337</td>\n",
       "      <td>-1.079591</td>\n",
       "      <td>-0.265196</td>\n",
       "      <td>-0.537930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-06-30</th>\n",
       "      <td>-1.197664</td>\n",
       "      <td>-0.473421</td>\n",
       "      <td>-0.944389</td>\n",
       "      <td>0.586640</td>\n",
       "      <td>1.965422</td>\n",
       "      <td>-0.431782</td>\n",
       "      <td>0.249136</td>\n",
       "      <td>2.101166</td>\n",
       "      <td>0.525071</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             traffic  paid_traffic  totalRevenue     likes   retweet  \\\n",
       "2019-03-31  1.286393      3.097198     -0.224631 -0.886782 -1.151780   \n",
       "2018-12-31  1.393636      1.920023      0.261087 -1.025690 -1.014412   \n",
       "2018-09-30  1.138647      0.528741      2.377258 -0.782600 -0.993278   \n",
       "2018-06-30  1.380319     -0.322606      0.680680  0.589121 -0.528339   \n",
       "2018-03-31  1.358805     -0.337641     -0.106696  1.338234 -0.031700   \n",
       "2017-12-31  0.397090     -0.410358     -0.268477  2.065023  1.489917   \n",
       "2017-09-30  0.149421     -0.291733      1.840531  1.241495  0.929877   \n",
       "2017-06-30  0.073507     -0.326479      0.017903  0.204642  0.221903   \n",
       "2017-03-31 -0.221000     -0.432830     -0.764636 -0.410524 -0.338137   \n",
       "2016-12-31 -0.617544     -0.507735     -0.856410 -0.256733 -0.443805   \n",
       "2016-09-30 -0.926877     -0.489698      0.879121  0.971119  1.743520   \n",
       "2016-06-30 -1.000752     -0.492452     -0.588126 -0.184798 -0.221903   \n",
       "2016-03-31 -1.067081     -0.481730     -1.115442 -1.211728 -0.813643   \n",
       "2015-12-31 -1.051574     -0.491558     -1.306985 -1.253897 -1.109513   \n",
       "2015-09-30 -1.095325     -0.487721      0.119213 -0.983522  0.295870   \n",
       "2015-06-30 -1.197664     -0.473421     -0.944389  0.586640  1.965422   \n",
       "\n",
       "            count_comments  senti_negative  senti_neutral  senti_positive  \n",
       "2019-03-31        0.320742       -0.415227      -1.081182       -0.743672  \n",
       "2018-12-31        0.179290       -1.079591      -0.917985       -0.846543  \n",
       "2018-09-30        0.886550       -1.079591      -0.917985       -0.846543  \n",
       "2018-06-30       -0.064007        0.581318      -0.673189        0.593652  \n",
       "2018-03-31       -0.035717        1.910046       1.121982        1.005136  \n",
       "2017-12-31        1.537230        1.577864       1.693173        2.479620  \n",
       "2017-09-30        2.550026        1.910046       1.529975        1.690942  \n",
       "2017-06-30        0.479169       -0.415227      -0.020400        0.490781  \n",
       "2017-03-31       -0.929693        0.249136      -0.509992       -0.435059  \n",
       "2016-12-31       -1.037197       -0.415227      -0.509992       -0.640801  \n",
       "2016-09-30        0.281136       -0.083045       0.224396       -0.195026  \n",
       "2016-06-30       -0.929693       -0.083045      -0.591590       -0.709381  \n",
       "2016-03-31       -1.206939       -1.079591      -0.836386       -1.017995  \n",
       "2015-12-31       -1.093778       -0.747409      -0.346794       -0.812252  \n",
       "2015-09-30       -0.505337       -1.079591      -0.265196       -0.537930  \n",
       "2015-06-30       -0.431782        0.249136       2.101166        0.525071  "
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore data\n",
    "webtraffic_tweet_fund_df_nom"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "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>traffic</th>\n",
       "      <th>paid_traffic</th>\n",
       "      <th>totalRevenue</th>\n",
       "      <th>likes</th>\n",
       "      <th>retweet</th>\n",
       "      <th>count_comments</th>\n",
       "      <th>senti_negative</th>\n",
       "      <th>senti_neutral</th>\n",
       "      <th>senti_positive</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>traffic</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.600971</td>\n",
       "      <td>0.486424</td>\n",
       "      <td>0.169449</td>\n",
       "      <td>-0.358751</td>\n",
       "      <td>0.474140</td>\n",
       "      <td>0.262173</td>\n",
       "      <td>-0.148611</td>\n",
       "      <td>0.227598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>paid_traffic</th>\n",
       "      <td>0.600971</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.165000</td>\n",
       "      <td>-0.357652</td>\n",
       "      <td>-0.457775</td>\n",
       "      <td>0.195441</td>\n",
       "      <td>-0.271918</td>\n",
       "      <td>-0.407725</td>\n",
       "      <td>-0.294768</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>totalRevenue</th>\n",
       "      <td>0.486424</td>\n",
       "      <td>0.165000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.242425</td>\n",
       "      <td>0.089407</td>\n",
       "      <td>0.740516</td>\n",
       "      <td>0.154957</td>\n",
       "      <td>0.021636</td>\n",
       "      <td>0.215627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>likes</th>\n",
       "      <td>0.169449</td>\n",
       "      <td>-0.357652</td>\n",
       "      <td>0.242425</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.749258</td>\n",
       "      <td>0.585070</td>\n",
       "      <td>0.884749</td>\n",
       "      <td>0.795292</td>\n",
       "      <td>0.911202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>retweet</th>\n",
       "      <td>-0.358751</td>\n",
       "      <td>-0.457775</td>\n",
       "      <td>0.089407</td>\n",
       "      <td>0.749258</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.377470</td>\n",
       "      <td>0.510280</td>\n",
       "      <td>0.836692</td>\n",
       "      <td>0.652346</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>count_comments</th>\n",
       "      <td>0.474140</td>\n",
       "      <td>0.195441</td>\n",
       "      <td>0.740516</td>\n",
       "      <td>0.585070</td>\n",
       "      <td>0.377470</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.532460</td>\n",
       "      <td>0.462752</td>\n",
       "      <td>0.686259</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>senti_negative</th>\n",
       "      <td>0.262173</td>\n",
       "      <td>-0.271918</td>\n",
       "      <td>0.154957</td>\n",
       "      <td>0.884749</td>\n",
       "      <td>0.510280</td>\n",
       "      <td>0.532460</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.742016</td>\n",
       "      <td>0.874409</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>senti_neutral</th>\n",
       "      <td>-0.148611</td>\n",
       "      <td>-0.407725</td>\n",
       "      <td>0.021636</td>\n",
       "      <td>0.795292</td>\n",
       "      <td>0.836692</td>\n",
       "      <td>0.462752</td>\n",
       "      <td>0.742016</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.822844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>senti_positive</th>\n",
       "      <td>0.227598</td>\n",
       "      <td>-0.294768</td>\n",
       "      <td>0.215627</td>\n",
       "      <td>0.911202</td>\n",
       "      <td>0.652346</td>\n",
       "      <td>0.686259</td>\n",
       "      <td>0.874409</td>\n",
       "      <td>0.822844</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 traffic  paid_traffic  totalRevenue     likes   retweet  \\\n",
       "traffic         1.000000      0.600971      0.486424  0.169449 -0.358751   \n",
       "paid_traffic    0.600971      1.000000      0.165000 -0.357652 -0.457775   \n",
       "totalRevenue    0.486424      0.165000      1.000000  0.242425  0.089407   \n",
       "likes           0.169449     -0.357652      0.242425  1.000000  0.749258   \n",
       "retweet        -0.358751     -0.457775      0.089407  0.749258  1.000000   \n",
       "count_comments  0.474140      0.195441      0.740516  0.585070  0.377470   \n",
       "senti_negative  0.262173     -0.271918      0.154957  0.884749  0.510280   \n",
       "senti_neutral  -0.148611     -0.407725      0.021636  0.795292  0.836692   \n",
       "senti_positive  0.227598     -0.294768      0.215627  0.911202  0.652346   \n",
       "\n",
       "                count_comments  senti_negative  senti_neutral  senti_positive  \n",
       "traffic               0.474140        0.262173      -0.148611        0.227598  \n",
       "paid_traffic          0.195441       -0.271918      -0.407725       -0.294768  \n",
       "totalRevenue          0.740516        0.154957       0.021636        0.215627  \n",
       "likes                 0.585070        0.884749       0.795292        0.911202  \n",
       "retweet               0.377470        0.510280       0.836692        0.652346  \n",
       "count_comments        1.000000        0.532460       0.462752        0.686259  \n",
       "senti_negative        0.532460        1.000000       0.742016        0.874409  \n",
       "senti_neutral         0.462752        0.742016       1.000000        0.822844  \n",
       "senti_positive        0.686259        0.874409       0.822844        1.000000  "
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Explore correlaration between Revenue and features\n",
    "webtraffic_tweet_fund_df_nom.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([<matplotlib.axes._subplots.AxesSubplot object at 0x0000017B964EA2B0>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B96508E80>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B965336D8>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B9655AF28>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B965C67B8>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B965ECFD0>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B9661B860>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B96649128>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x0000017B96649160>],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "webtraffic_tweet_fund_df_nom.plot(subplots=True, figsize=(10,6))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Prepare data for Stats and Machine Learning models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((16, 8), (16,))"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Select target(dependent/label) and feature(independent) variables\n",
    "targets = webtraffic_tweet_fund_df_nom['totalRevenue']\n",
    "features = webtraffic_tweet_fund_df_nom.drop(['totalRevenue'], axis=1)\n",
    "feature_names = features.columns\n",
    "features.shape, targets.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['const', 'traffic', 'paid_traffic', 'likes', 'retweet',\n",
       "       'count_comments', 'senti_negative', 'senti_neutral', 'senti_positive'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Using a linear OLS model here \n",
    "import statsmodels.api as sm\n",
    "# Add a constant to the features\n",
    "linear_features = sm.add_constant(features)\n",
    "linear_features.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(16, 9) (13, 9) (3, 9)\n"
     ]
    }
   ],
   "source": [
    "# Create a size for the training set that is 85% of the total number of samples\n",
    "train_size = int(0.85 * targets.shape[0])\n",
    "train_features = linear_features[:train_size]\n",
    "train_targets = targets[:train_size]\n",
    "test_features = linear_features[train_size:]\n",
    "test_targets = targets[train_size:]\n",
    "print(linear_features.shape, train_features.shape, test_features.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:           totalRevenue   R-squared:                       0.996\n",
      "Model:                            OLS   Adj. R-squared:                  0.987\n",
      "Method:                 Least Squares   F-statistic:                     111.7\n",
      "Date:                Mon, 17 Jun 2019   Prob (F-statistic):           0.000197\n",
      "Time:                        00:40:09   Log-Likelihood:                 16.722\n",
      "No. Observations:                  13   AIC:                            -15.44\n",
      "Df Residuals:                       4   BIC:                            -10.36\n",
      "Df Model:                           8                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==================================================================================\n",
      "                     coef    std err          t      P>|t|      [0.025      0.975]\n",
      "----------------------------------------------------------------------------------\n",
      "const             -0.1958      0.045     -4.331      0.012      -0.321      -0.070\n",
      "traffic            0.5248      0.106      4.949      0.008       0.230       0.819\n",
      "paid_traffic      -0.7196      0.060    -11.919      0.000      -0.887      -0.552\n",
      "likes              0.5933      0.233      2.550      0.063      -0.053       1.239\n",
      "retweet            0.0070      0.186      0.038      0.972      -0.509       0.523\n",
      "count_comments     1.5110      0.071     21.338      0.000       1.314       1.708\n",
      "senti_negative     0.2260      0.107      2.119      0.101      -0.070       0.522\n",
      "senti_neutral     -0.5232      0.133     -3.920      0.017      -0.894      -0.153\n",
      "senti_positive    -1.4640      0.117    -12.500      0.000      -1.789      -1.139\n",
      "==============================================================================\n",
      "Omnibus:                        2.022   Durbin-Watson:                   1.390\n",
      "Prob(Omnibus):                  0.364   Jarque-Bera (JB):                0.580\n",
      "Skew:                           0.492   Prob(JB):                        0.748\n",
      "Kurtosis:                       3.316   Cond. No.                         21.2\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\eche1\\Anaconda3\\lib\\site-packages\\scipy\\stats\\stats.py:1394: UserWarning: kurtosistest only valid for n>=20 ... continuing anyway, n=13\n",
      "  \"anyway, n=%i\" % int(n))\n"
     ]
    }
   ],
   "source": [
    "# Create the linear model and complete the least squares fit\n",
    "model = sm.OLS(train_targets, train_features)\n",
    "results = model.fit()  # fit the model\n",
    "print(results.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "const             0.012337\n",
      "traffic           0.007764\n",
      "paid_traffic      0.000284\n",
      "likes             0.063336\n",
      "retweet           0.971779\n",
      "count_comments    0.000029\n",
      "senti_negative    0.101496\n",
      "senti_neutral     0.017243\n",
      "senti_positive    0.000236\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# examine pvalues\n",
    "# Features with p <= 0.05 are typically considered significantly different from 0\n",
    "print(results.pvalues)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make predictions from our model for train and test sets\n",
    "train_predictions = results.predict(train_features)\n",
    "test_predictions = results.predict(test_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Scatter the predictions vs the targets with 80% transparency\n",
    "plt.scatter(train_predictions, train_targets, alpha=0.2, color='b', label='train')\n",
    "plt.scatter(test_predictions, test_targets, alpha=0.2, color='r', label='test')\n",
    "\n",
    "# Plot the perfect prediction line\n",
    "xmin, xmax = plt.xlim()\n",
    "plt.plot(np.arange(xmin, xmax, 0.01), np.arange(xmin, xmax, 0.01), c='k')\n",
    "\n",
    "# Set the axis labels and show the plot\n",
    "plt.xlabel('predictions')\n",
    "plt.ylabel('actual')\n",
    "plt.legend()  # show the legend\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.0\n",
      "-2.921474529396102\n"
     ]
    }
   ],
   "source": [
    "## trying Decision Tree Models\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "\n",
    "# Create a decision tree regression model with default arguments\n",
    "decision_tree = DecisionTreeRegressor()\n",
    "\n",
    "# Fit the model to the training features and targets\n",
    "decision_tree.fit(train_features, train_targets)\n",
    "\n",
    "# Check the score on train and test\n",
    "print(decision_tree.score(train_features, train_targets))\n",
    "print(decision_tree.score(test_features, test_targets))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8580535745035016\n",
      "-1.2073638174918648\n"
     ]
    }
   ],
   "source": [
    "# Try with Random Forest\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "\n",
    "# Create the random forest model and fit to the training data\n",
    "rfr = RandomForestRegressor(n_estimators=200)\n",
    "rfr.fit(train_features, train_targets)\n",
    "\n",
    "# Look at the R^2 scores on train and test\n",
    "print(rfr.score(train_features, train_targets))\n",
    "print(rfr.score(test_features, test_targets))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Get feature importances from our random forest model\n",
    "importances = rfr.feature_importances_\n",
    "\n",
    "importances = importances[1:]\n",
    "\n",
    "# Get the index of importances from greatest importance to least\n",
    "sorted_index = np.argsort(importances)[::-1]\n",
    "x = range(len(importances))\n",
    "\n",
    "# Create tick labels \n",
    "labels = np.array(feature_names)[sorted_index]\n",
    "plt.bar(x, importances[sorted_index], tick_label=labels)\n",
    "\n",
    "# Rotate tick labels to vertical\n",
    "plt.xticks(rotation=90)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}