ASDRP_Sentiment_Analysis_Stock_Price_Prediction.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "ASDRP Sentiment Analysis - Stock Price Prediction",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nJ6kpvxIEf-o",
        "outputId": "b7c8ad92-29fe-4118-f221-df6380f9fa12"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sLKbYbTTvO2o",
        "outputId": "06582ee2-cb24-491d-c70d-4f323a525751"
      },
      "source": [
        "%ls"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[0m\u001b[01;34mdrive\u001b[0m/  \u001b[01;34msample_data\u001b[0m/\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8Nw8y29HCsxK"
      },
      "source": [
        "import io\n",
        "import re\n",
        "import collections\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "from datetime import datetime\n",
        "import progressbar\n",
        "import math"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 203
        },
        "id": "UHJl1grFAs92",
        "outputId": "604f9b0d-5ab7-476d-cbd4-edc9c6530ab7"
      },
      "source": [
        "# Read csv containing stocks and prices each day\n",
        "path = \"./drive/MyDrive/stock_prices.csv\"\n",
        "df_stock_prices = pd.read_csv(path)\n",
        "df_stock_prices.columns= df_stock_prices.columns.str.lower()\n",
        "df_stock_prices.rename(columns={'adj close': 'adj_close'}, inplace=True)\n",
        "df_stock_prices['date'] = pd.to_datetime(df_stock_prices['date']).dt.date\n",
        "df_stock_prices.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>unnamed: 0</th>\n",
              "      <th>ticker</th>\n",
              "      <th>date</th>\n",
              "      <th>high</th>\n",
              "      <th>low</th>\n",
              "      <th>open</th>\n",
              "      <th>close</th>\n",
              "      <th>volume</th>\n",
              "      <th>adj_close</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>A</td>\n",
              "      <td>2010-01-04</td>\n",
              "      <td>22.625179</td>\n",
              "      <td>22.267525</td>\n",
              "      <td>22.453505</td>\n",
              "      <td>22.389128</td>\n",
              "      <td>3815561</td>\n",
              "      <td>20.487324</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>A</td>\n",
              "      <td>2010-01-05</td>\n",
              "      <td>22.331903</td>\n",
              "      <td>22.002861</td>\n",
              "      <td>22.324751</td>\n",
              "      <td>22.145924</td>\n",
              "      <td>4186031</td>\n",
              "      <td>20.264782</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>A</td>\n",
              "      <td>2010-01-06</td>\n",
              "      <td>22.174536</td>\n",
              "      <td>22.002861</td>\n",
              "      <td>22.067240</td>\n",
              "      <td>22.067240</td>\n",
              "      <td>3243779</td>\n",
              "      <td>20.192783</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>A</td>\n",
              "      <td>2010-01-07</td>\n",
              "      <td>22.045780</td>\n",
              "      <td>21.816881</td>\n",
              "      <td>22.017166</td>\n",
              "      <td>22.038628</td>\n",
              "      <td>3095172</td>\n",
              "      <td>20.166595</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>A</td>\n",
              "      <td>2010-01-08</td>\n",
              "      <td>22.067240</td>\n",
              "      <td>21.745352</td>\n",
              "      <td>21.917025</td>\n",
              "      <td>22.031473</td>\n",
              "      <td>3733918</td>\n",
              "      <td>20.160051</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   unnamed: 0 ticker        date  ...      close   volume  adj_close\n",
              "0           0      A  2010-01-04  ...  22.389128  3815561  20.487324\n",
              "1           1      A  2010-01-05  ...  22.145924  4186031  20.264782\n",
              "2           2      A  2010-01-06  ...  22.067240  3243779  20.192783\n",
              "3           3      A  2010-01-07  ...  22.038628  3095172  20.166595\n",
              "4           4      A  2010-01-08  ...  22.031473  3733918  20.160051\n",
              "\n",
              "[5 rows x 9 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZQxRYHa3X6ff",
        "outputId": "50786344-7c93-40f9-9049-813e320f609f"
      },
      "source": [
        "# Get all stock tickers\n",
        "tickers = df_stock_prices['ticker'].unique()\n",
        "tickers"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array(['A', 'AAL', 'AAP', 'AAPL', 'ABBV', 'ABC', 'ABMD', 'ABT', 'ACN',\n",
              "       'ADBE', 'ADI', 'ADM', 'ADP', 'ADSK', 'AEE', 'AEP', 'AES', 'AFL',\n",
              "       'AIG', 'AIZ', 'AJG', 'AKAM', 'ALB', 'ALGN', 'ALK', 'ALL', 'ALLE',\n",
              "       'AMAT', 'AMCR', 'AMD', 'AME', 'AMGN', 'AMP', 'AMT', 'AMZN', 'ANET',\n",
              "       'ANSS', 'ANTM', 'AON', 'AOS', 'APA', 'APD', 'APH', 'APTV', 'ARE',\n",
              "       'ATO', 'ATVI', 'AVB', 'AVGO', 'AVY', 'AWK', 'AXP', 'AZO', 'BA',\n",
              "       'BAC', 'BAX', 'BBY', 'BDX', 'BEN', 'BF-B', 'BIIB', 'BIO', 'BK',\n",
              "       'BKNG', 'BKR', 'BLK', 'BLL', 'BMY', 'BR', 'BRK-B', 'BSX', 'BWA',\n",
              "       'BXP', 'C', 'CAG', 'CAH', 'CARR', 'CAT', 'CB', 'CBOE', 'CBRE',\n",
              "       'CCI', 'CCL', 'CDNS', 'CDW', 'CE', 'CERN', 'CF', 'CFG', 'CHD',\n",
              "       'CHRW', 'CHTR', 'CI', 'CINF', 'CL', 'CLX', 'CMA', 'CMCSA', 'CME',\n",
              "       'CMG', 'CMI', 'CMS', 'CNC', 'CNP', 'COF', 'COG', 'COO', 'COP',\n",
              "       'COST', 'CPB', 'CPRT', 'CRL', 'CRM', 'CSCO', 'CSX', 'CTAS', 'CTLT',\n",
              "       'CTSH', 'CTVA', 'CTXS', 'CVS', 'CVX', 'CZR', 'D', 'DAL', 'DD',\n",
              "       'DE', 'DFS', 'DG', 'DGX', 'DHI', 'DHR', 'DIS', 'DISCA', 'DISCK',\n",
              "       'DISH', 'DLR', 'DLTR', 'DOV', 'DOW', 'DPZ', 'DRE', 'DRI', 'DTE',\n",
              "       'DUK', 'DVA', 'DVN', 'DXC', 'DXCM', 'EA', 'EBAY', 'ECL', 'ED',\n",
              "       'EFX', 'EIX', 'EL', 'EMN', 'EMR', 'ENPH', 'EOG', 'EQIX', 'EQR',\n",
              "       'ES', 'ESS', 'ETN', 'ETR', 'ETSY', 'EVRG', 'EW', 'EXC', 'EXPD',\n",
              "       'EXPE', 'EXR', 'F', 'FANG', 'FAST', 'FB', 'FBHS', 'FCX', 'FDX',\n",
              "       'FE', 'FFIV', 'FIS', 'FISV', 'FITB', 'FLT', 'FMC', 'FOX', 'FOXA',\n",
              "       'FRC', 'FRT', 'FTNT', 'FTV', 'GD', 'GE', 'GILD', 'GIS', 'GL',\n",
              "       'GLW', 'GM', 'GNRC', 'GOOG', 'GOOGL', 'GPC', 'GPN', 'GPS', 'GRMN',\n",
              "       'GS', 'GWW', 'HAL', 'HAS', 'HBAN', 'HBI', 'HCA', 'HD', 'HES',\n",
              "       'HIG', 'HII', 'HLT', 'HOLX', 'HON', 'HPE', 'HPQ', 'HRL', 'HSIC',\n",
              "       'HST', 'HSY', 'HUM', 'HWM', 'IBM', 'ICE', 'IDXX', 'IEX', 'IFF',\n",
              "       'ILMN', 'INCY', 'INFO', 'INTC', 'INTU', 'IP', 'IPG', 'IPGP', 'IQV',\n",
              "       'IR', 'IRM', 'ISRG', 'IT', 'ITW', 'IVZ', 'J', 'JBHT', 'JCI',\n",
              "       'JKHY', 'JNJ', 'JNPR', 'JPM', 'K', 'KEY', 'KEYS', 'KHC', 'KIM',\n",
              "       'KLAC', 'KMB', 'KMI', 'KMX', 'KO', 'KR', 'KSU', 'L', 'LB', 'LDOS',\n",
              "       'LEG', 'LEN', 'LH', 'LHX', 'LIN', 'LKQ', 'LLY', 'LMT', 'LNC',\n",
              "       'LNT', 'LOW', 'LRCX', 'LUMN', 'LUV', 'LVS', 'LW', 'LYB', 'LYV',\n",
              "       'MA', 'MAA', 'MAR', 'MAS', 'MCD', 'MCHP', 'MCK', 'MCO', 'MDLZ',\n",
              "       'MDT', 'MET', 'MGM', 'MHK', 'MKC', 'MKTX', 'MLM', 'MMC', 'MMM',\n",
              "       'MNST', 'MO', 'MOS', 'MPC', 'MPWR', 'MRK', 'MRNA', 'MRO', 'MS',\n",
              "       'MSCI', 'MSFT', 'MSI', 'MTB', 'MTD', 'MU', 'MXIM', 'NCLH', 'NDAQ',\n",
              "       'NEE', 'NEM', 'NFLX', 'NI', 'NKE', 'NLOK', 'NLSN', 'NOC', 'NOV',\n",
              "       'NOW', 'NRG', 'NSC', 'NTAP', 'NTRS', 'NUE', 'NVDA', 'NVR', 'NWL',\n",
              "       'NWS', 'NWSA', 'NXPI', 'O', 'ODFL', 'OGN', 'OKE', 'OMC', 'ORCL',\n",
              "       'ORLY', 'OTIS', 'OXY', 'PAYC', 'PAYX', 'PBCT', 'PCAR', 'PEAK',\n",
              "       'PEG', 'PENN', 'PEP', 'PFE', 'PFG', 'PG', 'PGR', 'PH', 'PHM',\n",
              "       'PKG', 'PKI', 'PLD', 'PM', 'PNC', 'PNR', 'PNW', 'POOL', 'PPG',\n",
              "       'PPL', 'PRGO', 'PRU', 'PSA', 'PSX', 'PTC', 'PVH', 'PWR', 'PXD',\n",
              "       'PYPL', 'QCOM', 'QRVO', 'RCL', 'RE', 'REG', 'REGN', 'RF', 'RHI',\n",
              "       'RJF', 'RL', 'RMD', 'ROK', 'ROL', 'ROP', 'ROST', 'RSG', 'RTX',\n",
              "       'SBAC', 'SBUX', 'SCHW', 'SEE', 'SHW', 'SIVB', 'SJM', 'SLB', 'SNA',\n",
              "       'SNPS', 'SO', 'SPG', 'SPGI', 'SRE', 'STE', 'STT', 'STX', 'STZ',\n",
              "       'SWK', 'SWKS', 'SYF', 'SYK', 'SYY', 'T', 'TAP', 'TDG', 'TDY',\n",
              "       'TEL', 'TER', 'TFC', 'TFX', 'TGT', 'TJX', 'TMO', 'TMUS', 'TPR',\n",
              "       'TRMB', 'TROW', 'TRV', 'TSCO', 'TSLA', 'TSN', 'TT', 'TTWO', 'TWTR',\n",
              "       'TXN', 'TXT', 'TYL', 'UA', 'UAA', 'UAL', 'UDR', 'UHS', 'ULTA',\n",
              "       'UNH', 'UNM', 'UNP', 'UPS', 'URI', 'USB', 'V', 'VFC', 'VIAC',\n",
              "       'VLO', 'VMC', 'VNO', 'VRSK', 'VRSN', 'VRTX', 'VTR', 'VTRS', 'VZ',\n",
              "       'WAB', 'WAT', 'WBA', 'WDC', 'WEC', 'WELL', 'WFC', 'WHR', 'WLTW',\n",
              "       'WM', 'WMB', 'WMT', 'WRB', 'WRK', 'WST', 'WU', 'WY', 'WYNN', 'XEL',\n",
              "       'XLNX', 'XOM', 'XRAY', 'XYL', 'YUM', 'ZBH', 'ZBRA', 'ZION', 'ZTS'],\n",
              "      dtype=object)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 325
        },
        "id": "hPdy-8cEBAAi",
        "outputId": "ca13dcbd-f448-4fe6-ebb0-1feda9331f45"
      },
      "source": [
        "# Plot average stock price over time for a specific stock\n",
        "ticker = 'T'\n",
        "\n",
        "price_ticker_mask = (df_stock_prices['ticker'] == ticker)\n",
        "x = df_stock_prices.loc[price_ticker_mask, 'date']\n",
        "y = df_stock_prices.loc[price_ticker_mask, 'adj_close']\n",
        "plt.xticks(rotation = 30)\n",
        "plt.plot(x, y)\n",
        "plt.fill_between(x,y, alpha=0.4)\n",
        "plt.xlabel(\"Date\", labelpad=13, fontsize=13)\n",
        "plt.ylabel(\"Adjusted Close Price\", labelpad=13, fontsize=13)\n",
        "plt.title(f\"{ticker} Stock Price\", y=1.02, fontsize=15);\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M_JsC3T-cf_G"
      },
      "source": [
        "# Removes special characters and punctuation from tweet\n",
        "# Code from https://www.geeksforgeeks.org/twitter-sentiment-analysis-using-python/\n",
        "def clean_tweet(tweet):\n",
        "    return ' '.join(re.sub(\"(@[A-Za-z0-9]+)|([^0-9A-Za-z \\t])|(\\w+:\\/\\/\\S+)\", \" \", tweet).split())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 417
        },
        "id": "AlBqA4IlFGea",
        "outputId": "cd911ee3-723c-4e7e-8e58-cc2854050c7b"
      },
      "source": [
        "# Read csv containaing stock tweets and sentiments\n",
        "path = \"./drive/MyDrive/stock_tweet_sentiments.csv\"\n",
        "df_tweet_sentiments = pd.read_csv(path)\n",
        "df_tweet_sentiments.columns= df_tweet_sentiments.columns.str.lower()\n",
        "df_tweet_sentiments['text'] = df_tweet_sentiments['text'].apply(clean_tweet)\n",
        "df_tweet_sentiments"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>text</th>\n",
              "      <th>sentiment</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Kickers on my watchlist XIDE TIT SOQ PNK CPW B...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>user AAP MOVIE 55 return for the FEA GEED indi...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>user I d be afraid to short AMZN they are look...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>MNTA Over 12 00</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>OI Over 21 37</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5786</th>\n",
              "      <td>Industry body CII said discoms are likely to s...</td>\n",
              "      <td>-1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5787</th>\n",
              "      <td>Gold prices slip below Rs 46 000 as investors ...</td>\n",
              "      <td>-1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5788</th>\n",
              "      <td>Workers at Bajaj Auto have agreed to a 10 wage...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5789</th>\n",
              "      <td>Sharemarket LIVE Sensex off day s high up 600 ...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5790</th>\n",
              "      <td>Sensex Nifty climb off day s highs still up 2 ...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5791 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                   text  sentiment\n",
              "0     Kickers on my watchlist XIDE TIT SOQ PNK CPW B...          1\n",
              "1     user AAP MOVIE 55 return for the FEA GEED indi...          1\n",
              "2     user I d be afraid to short AMZN they are look...          1\n",
              "3                                       MNTA Over 12 00          1\n",
              "4                                         OI Over 21 37          1\n",
              "...                                                 ...        ...\n",
              "5786  Industry body CII said discoms are likely to s...         -1\n",
              "5787  Gold prices slip below Rs 46 000 as investors ...         -1\n",
              "5788  Workers at Bajaj Auto have agreed to a 10 wage...          1\n",
              "5789  Sharemarket LIVE Sensex off day s high up 600 ...          1\n",
              "5790  Sensex Nifty climb off day s highs still up 2 ...          1\n",
              "\n",
              "[5791 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RWanrkkKiEVD",
        "outputId": "01a6d3e5-2d7d-4bd1-c30b-ec79040ae427"
      },
      "source": [
        "# Import nltk and metric libraries\n",
        "import nltk\n",
        "nltk.download('brown')\n",
        "nltk.download('punkt')\n",
        "!pip install vaderSentiment\n",
        "from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer\n",
        "\n",
        "from sklearn.metrics import accuracy_score, classification_report"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[nltk_data] Downloading package brown to /root/nltk_data...\n",
            "[nltk_data]   Package brown is already up-to-date!\n",
            "[nltk_data] Downloading package punkt to /root/nltk_data...\n",
            "[nltk_data]   Package punkt is already up-to-date!\n",
            "Requirement already satisfied: vaderSentiment in /usr/local/lib/python3.7/dist-packages (3.3.2)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.7/dist-packages (from vaderSentiment) (2.23.0)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests->vaderSentiment) (2021.5.30)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests->vaderSentiment) (1.24.3)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests->vaderSentiment) (3.0.4)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests->vaderSentiment) (2.10)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ssuxXwk0is3m",
        "outputId": "d3639ee9-aeac-4bb3-f289-42027d7fdff5"
      },
      "source": [
        "# Test Vader classifier\n",
        "sid = SentimentIntensityAnalyzer()\n",
        "\n",
        "# Calculate accuracy of Vader model from labelled tweets\n",
        "y_pred = []\n",
        "y_test = []\n",
        "count = 0\n",
        "for _, tweet in df_tweet_sentiments.iterrows():\n",
        "  polarity_score = sid.polarity_scores(tweet.text)[\"compound\"]\n",
        "  sentiment_pred = 1 if polarity_score >= 0 else -1\n",
        "  y_pred.append(sentiment_pred)\n",
        "  y_test.append(tweet.sentiment)\n",
        "\n",
        "accuracy = round(accuracy_score(y_test, y_pred) * 100, 2)\n",
        "print(f\"Accuracy is {accuracy}%\")\n",
        "print(\"Classification Report:\\n\")\n",
        "print(classification_report(y_test, y_pred))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Accuracy is 66.4%\n",
            "Classification Report:\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "          -1       0.56      0.36      0.43      2106\n",
            "           1       0.70      0.84      0.76      3685\n",
            "\n",
            "    accuracy                           0.66      5791\n",
            "   macro avg       0.63      0.60      0.60      5791\n",
            "weighted avg       0.65      0.66      0.64      5791\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4hyNIH31gR6E",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 434
        },
        "outputId": "df37fed6-c5c3-4b58-d845-85410eea4f53"
      },
      "source": [
        "# Get tweets about top companies from 2015 to 2020\n",
        "path = \"./drive/MyDrive/stock_tweets.csv\"\n",
        "df_stock_tweets = pd.read_csv(path)\n",
        "\n",
        "df_stock_tweets.drop(columns=['tweet_id', 'writer', 'comment_num', 'retweet_num', 'like_num'], inplace=True)\n",
        "df_stock_tweets.rename(columns={'post_date': 'date', 'body': 'text'}, inplace=True)\n",
        "df_stock_tweets['date'] = pd.to_datetime(df_stock_tweets['date'], unit='s').dt.date\n",
        "df_stock_tweets['text'] = df_stock_tweets['text'].apply(clean_tweet)\n",
        "df_stock_tweets.sort_values(by='date', inplace=True)\n",
        "print(df_stock_tweets.shape[0])\n",
        "df_stock_tweets"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "3717964\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "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",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2015-01-01</td>\n",
              "      <td>lx21 made 10 008 on AAPL Check it out Learn ho...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>386</th>\n",
              "      <td>2015-01-01</td>\n",
              "      <td>Google Fiber s latest FCC filing is Comcast s ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>387</th>\n",
              "      <td>2015-01-01</td>\n",
              "      <td>Jeff Bezos lost 7 4 billion in Amazon s worst ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>388</th>\n",
              "      <td>2015-01-01</td>\n",
              "      <td>TSLA Weekly Updated via</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>389</th>\n",
              "      <td>2015-01-01</td>\n",
              "      <td>MSFT What Microsoft s fresh start browser stra...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3716509</th>\n",
              "      <td>2019-12-31</td>\n",
              "      <td>Stock Market Watchlist for 31 Dec 2019 SPY BA ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3716508</th>\n",
              "      <td>2019-12-31</td>\n",
              "      <td>AMZN ZNGA Amazon Zynga Named Top Internet Sect...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3716507</th>\n",
              "      <td>2019-12-31</td>\n",
              "      <td>Trading Seasonality using Vertical Call Spread...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3716521</th>\n",
              "      <td>2019-12-31</td>\n",
              "      <td>Good morning folks last trading day of the dec...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3717963</th>\n",
              "      <td>2019-12-31</td>\n",
              "      <td>AAPL patent 10 522 475 Vertical interconnects ...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>3717964 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "               date                                               text\n",
              "0        2015-01-01  lx21 made 10 008 on AAPL Check it out Learn ho...\n",
              "386      2015-01-01  Google Fiber s latest FCC filing is Comcast s ...\n",
              "387      2015-01-01  Jeff Bezos lost 7 4 billion in Amazon s worst ...\n",
              "388      2015-01-01                            TSLA Weekly Updated via\n",
              "389      2015-01-01  MSFT What Microsoft s fresh start browser stra...\n",
              "...             ...                                                ...\n",
              "3716509  2019-12-31  Stock Market Watchlist for 31 Dec 2019 SPY BA ...\n",
              "3716508  2019-12-31  AMZN ZNGA Amazon Zynga Named Top Internet Sect...\n",
              "3716507  2019-12-31  Trading Seasonality using Vertical Call Spread...\n",
              "3716521  2019-12-31  Good morning folks last trading day of the dec...\n",
              "3717963  2019-12-31  AAPL patent 10 522 475 Vertical interconnects ...\n",
              "\n",
              "[3717964 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ItgWmk1LGNP8",
        "outputId": "d810cd91-124b-473d-fba5-e6c8b1c487be"
      },
      "source": [
        "# Remove dates from stock price dataframe that aren't included in the stock tweets dataset, and only keep the 1 million most recent tweets\n",
        "num_tweets = 2000000\n",
        "\n",
        "df_stock_tweets = df_stock_tweets.tail(num_tweets) # Reduce dataset size, select starting from most recent tweets\n",
        "df_stock_tweets.reset_index(drop=True, inplace=True)\n",
        "\n",
        "start_date = df_stock_tweets.iloc[0]['date']\n",
        "end_date = df_stock_tweets.iloc[-1]['date']\n",
        "\n",
        "stock_price_mask = (df_stock_prices['date'] >= start_date) & (df_stock_prices['date'] <= end_date)\n",
        "df_stock_prices = df_stock_prices.loc[stock_price_mask]\n",
        "\n",
        "print(df_stock_prices.shape[0], df_stock_tweets.shape[0])\n",
        "print(df_stock_prices)\n",
        "print(df_stock_tweets)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "350877 2000000\n",
            "         unnamed: 0 ticker        date  ...       close   volume   adj_close\n",
            "1812           1812      A  2017-03-16  ...   53.049999  2676900   51.091694\n",
            "1813           1813      A  2017-03-17  ...   53.310001  1936900   51.342113\n",
            "1814           1814      A  2017-03-20  ...   53.150002  1221400   51.188011\n",
            "1815           1815      A  2017-03-21  ...   52.660000  1492700   50.716099\n",
            "1816           1816      A  2017-03-22  ...   53.130001  1257100   51.168747\n",
            "...             ...    ...         ...  ...         ...      ...         ...\n",
            "1409312     1409312    ZTS  2019-12-24  ...  132.919998   442500  131.595566\n",
            "1409313     1409313    ZTS  2019-12-26  ...  133.029999   929400  131.704453\n",
            "1409314     1409314    ZTS  2019-12-27  ...  133.250000  1296100  131.922272\n",
            "1409315     1409315    ZTS  2019-12-30  ...  132.210007   942000  130.892654\n",
            "1409316     1409316    ZTS  2019-12-31  ...  132.350006  1128600  131.031250\n",
            "\n",
            "[350877 rows x 9 columns]\n",
            "               date                                               text\n",
            "0        2017-03-16  Apple Inc AAPL Sets New 12 Month High After An...\n",
            "1        2017-03-16  I m no expert but aren t convertible notes usu...\n",
            "2        2017-03-16    Wonders shall never end TSLA rises on offerings\n",
            "3        2017-03-16  TSLA AAPL Stocks Back Near Record Highs These ...\n",
            "4        2017-03-16  AAPL COMPQ NQ F From September 2016 152 02 tar...\n",
            "...             ...                                                ...\n",
            "1999995  2019-12-31  Stock Market Watchlist for 31 Dec 2019 SPY BA ...\n",
            "1999996  2019-12-31  AMZN ZNGA Amazon Zynga Named Top Internet Sect...\n",
            "1999997  2019-12-31  Trading Seasonality using Vertical Call Spread...\n",
            "1999998  2019-12-31  Good morning folks last trading day of the dec...\n",
            "1999999  2019-12-31  AAPL patent 10 522 475 Vertical interconnects ...\n",
            "\n",
            "[2000000 rows x 2 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wKS9Dd70hzQ8"
      },
      "source": [
        "# Returns all stock mentioned in a tweet\n",
        "def get_stocks_in_tweet(tweet):\n",
        "  tweet = tweet.split()\n",
        "  return [ticker for ticker in tickers if ticker in tweet]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 450
        },
        "id": "bbcDRwtv5VXH",
        "outputId": "82e17a48-2349-4864-a808-9f8b06377345"
      },
      "source": [
        "# Get average sentiment each day for all mentioned stocks\n",
        "# stock_sentiment_dict = collections.defaultdict(list)\n",
        "stock_sentiment_dict = dict()\n",
        "\n",
        "start_time = datetime.now()\n",
        "\n",
        "# Create progress bar\n",
        "widgets = [\n",
        "  ' [', progressbar.Timer(), '] ',\n",
        "  progressbar.Bar(),\n",
        "  ' (', progressbar.ETA(), ') '\n",
        "]\n",
        "bar = progressbar.ProgressBar(\n",
        "    widgets=widgets, max_value=df_stock_tweets.shape[0]).start()\n",
        "\n",
        "for index, tweet in df_stock_tweets.iterrows():\n",
        "  # Update progress bar\n",
        "  bar.update(index + 1)\n",
        "\n",
        "  date = tweet['date']\n",
        "  polarity_score = sid.polarity_scores(tweet['text'])['compound']\n",
        "  mentioned_stocks = get_stocks_in_tweet(tweet['text'])\n",
        "  for stock in mentioned_stocks:\n",
        "    if (stock, date) in stock_sentiment_dict:\n",
        "      sum_polarity_score, len_polarity_score = stock_sentiment_dict[(stock, date)]\n",
        "      stock_sentiment_dict[(stock, date)] = (sum_polarity_score+polarity_score, len_polarity_score+1)\n",
        "    else:\n",
        "      stock_sentiment_dict[(stock, date)] = (polarity_score, 1)\n",
        "    # stock_sentiment_dict[(stock, date)].append(polarity_score)\n",
        "    \n",
        "bar.finish()\n",
        "\n",
        "stock_sentiments = list()\n",
        "for (stock, date), (sum_polarity_score, len_polarity_score) in stock_sentiment_dict.items():\n",
        "  # avg_polarity_score = sum(polarity_scores)/len(polarity_scores)\n",
        "  stock_sentiments.append([stock, date, round(sum_polarity_score/len_polarity_score, 4)])\n",
        "\n",
        "df_stock_sentiments = pd.DataFrame(stock_sentiments,columns=['ticker','date','avg_sentiment'])\n",
        "df_stock_sentiments.sort_values(by=['ticker', 'date'], inplace=True)\n",
        "\n",
        "\n",
        "end_time = datetime.now()\n",
        "print(f\"Runtime: {str(end_time - start_time)[:-3]}\")\n",
        "df_stock_sentiments.loc[df_stock_sentiments['ticker']=='TSLA']"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            " [Elapsed Time: 0:18:41] |##################################| (Time:  0:18:41) \n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Runtime: 0:18:42.390\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "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>ticker</th>\n",
              "      <th>date</th>\n",
              "      <th>avg_sentiment</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2017-03-16</td>\n",
              "      <td>0.0873</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>48</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2017-03-17</td>\n",
              "      <td>0.1076</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>211</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2017-03-18</td>\n",
              "      <td>0.1731</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>312</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2017-03-19</td>\n",
              "      <td>0.2035</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>395</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2017-03-20</td>\n",
              "      <td>0.1061</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>189628</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2019-12-27</td>\n",
              "      <td>0.1169</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>189860</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2019-12-28</td>\n",
              "      <td>0.1754</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>190000</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2019-12-29</td>\n",
              "      <td>0.1394</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>190161</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2019-12-30</td>\n",
              "      <td>0.1223</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>190375</th>\n",
              "      <td>TSLA</td>\n",
              "      <td>2019-12-31</td>\n",
              "      <td>0.1475</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>1021 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       ticker        date  avg_sentiment\n",
              "1        TSLA  2017-03-16         0.0873\n",
              "48       TSLA  2017-03-17         0.1076\n",
              "211      TSLA  2017-03-18         0.1731\n",
              "312      TSLA  2017-03-19         0.2035\n",
              "395      TSLA  2017-03-20         0.1061\n",
              "...       ...         ...            ...\n",
              "189628   TSLA  2019-12-27         0.1169\n",
              "189860   TSLA  2019-12-28         0.1754\n",
              "190000   TSLA  2019-12-29         0.1394\n",
              "190161   TSLA  2019-12-30         0.1223\n",
              "190375   TSLA  2019-12-31         0.1475\n",
              "\n",
              "[1021 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "je0FrY_cmSBH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 333
        },
        "outputId": "e61b35dd-9250-4e80-e9fc-b8fcf5ea50a0"
      },
      "source": [
        "# Plot average sentiment over time for a specific stock\n",
        "ticker = 'CVS'\n",
        "\n",
        "sentiment_date_mask = (df_stock_sentiments['date'] >= start_date) & (df_stock_sentiments['date'] <= end_date)\n",
        "sentiment_ticker_mask = (df_stock_sentiments['ticker'] == ticker)\n",
        "x = df_stock_sentiments.loc[sentiment_date_mask & sentiment_ticker_mask, 'date']\n",
        "y = df_stock_sentiments.loc[sentiment_date_mask & sentiment_ticker_mask, 'avg_sentiment']\n",
        "plt.xticks(rotation=30)\n",
        "plt.plot(x, y)\n",
        "plt.xlabel(\"Date\", labelpad=13, fontsize=13)\n",
        "plt.ylabel(\"Average Sentiment\", labelpad=13, fontsize=13)\n",
        "plt.title(f\"{ticker} Sentiment\", y=1.02, fontsize=15);\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "jsjUET8bkJOG",
        "outputId": "1bea151e-f115-4d0b-f5b0-3acf3317ea81"
      },
      "source": [
        "# Merge stock prices and stock sentiments by date and ticker\n",
        "df_merged = pd.merge(df_stock_prices[['ticker', 'date', 'adj_close']], df_stock_sentiments[['ticker', 'date', 'avg_sentiment']], how='inner')\n",
        "df_merged = df_merged.loc[df_merged['ticker'] == ticker, ['date', 'adj_close', 'avg_sentiment']]\n",
        "df_merged.set_index('date', inplace=True)\n",
        "df_merged"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>adj_close</th>\n",
              "      <th>avg_sentiment</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>date</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2017-03-22</th>\n",
              "      <td>68.767570</td>\n",
              "      <td>0.0000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2017-03-23</th>\n",
              "      <td>68.627342</td>\n",
              "      <td>0.5106</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2017-03-28</th>\n",
              "      <td>68.899040</td>\n",
              "      <td>-0.4404</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2017-03-29</th>\n",
              "      <td>69.030502</td>\n",
              "      <td>0.8316</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2017-03-30</th>\n",
              "      <td>69.109406</td>\n",
              "      <td>0.8316</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-23</th>\n",
              "      <td>70.714760</td>\n",
              "      <td>-0.0427</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-24</th>\n",
              "      <td>70.838348</td>\n",
              "      <td>-0.0516</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-26</th>\n",
              "      <td>70.809837</td>\n",
              "      <td>0.1808</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-30</th>\n",
              "      <td>70.068268</td>\n",
              "      <td>0.4421</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-31</th>\n",
              "      <td>70.629189</td>\n",
              "      <td>0.2194</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>644 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "            adj_close  avg_sentiment\n",
              "date                                \n",
              "2017-03-22  68.767570         0.0000\n",
              "2017-03-23  68.627342         0.5106\n",
              "2017-03-28  68.899040        -0.4404\n",
              "2017-03-29  69.030502         0.8316\n",
              "2017-03-30  69.109406         0.8316\n",
              "...               ...            ...\n",
              "2019-12-23  70.714760        -0.0427\n",
              "2019-12-24  70.838348        -0.0516\n",
              "2019-12-26  70.809837         0.1808\n",
              "2019-12-30  70.068268         0.4421\n",
              "2019-12-31  70.629189         0.2194\n",
              "\n",
              "[644 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 123
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 407
        },
        "id": "ZTEgvIgIK-OE",
        "outputId": "fe0df57f-2eb7-45ad-afeb-8011b138ffae"
      },
      "source": [
        "# Plot adjacent closing price and average sentiment over time\n",
        "# Code from https://towardsdatascience.com/granger-causality-and-vector-auto-regressive-model-for-time-series-forecasting-3226a64889a6\n",
        "fig, axes = plt.subplots(nrows=2, ncols=1, dpi=100)\n",
        "for i, ax in enumerate(axes.flatten()):\n",
        "  data = df_merged[df_merged.columns[i]]\n",
        "  ax.plot(data, color='red', linewidth=1)\n",
        "  ax.set_title(df_merged.columns[i])\n",
        "  ax.xaxis.set_ticks_position('none')\n",
        "  ax.yaxis.set_ticks_position('none')\n",
        "  ax.spines['top'].set_alpha(0)\n",
        "  ax.tick_params(labelsize=6)\n",
        "plt.tight_layout();"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 600x400 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SzNxYeB0ZhWH",
        "outputId": "ded59970-b681-46c2-bc98-037979213f0a"
      },
      "source": [
        "# Test adjusted closing prices and average sentiment for normality and skew\n",
        "from scipy import stats\n",
        "stat_price, p_price = stats.normaltest(df_merged['adj_close'])\n",
        "stat_sentiment, p_sentiment = stats.normaltest(df_merged['avg_sentiment'])\n",
        "alpha = 0.05\n",
        "print(\"Stock normality statistic=%.3f, p=%.3f\" % (stat_price, p_price))\n",
        "print(\"Sentiment normality statistic=%.3f, p=%.3f\" % (stat_sentiment, p_sentiment))\n",
        "print()\n",
        "\n",
        "print(\"Stock price data looks Gaussian\" if p_price > alpha else \"Stock price data looks non-Gaussian\")\n",
        "print(\"Stock sentiment data looks Gaussian\" if p_sentiment > alpha else \"Stock sentiment data looks non-Gaussian\")\n",
        "print()\n",
        "\n",
        "print(f\"Stock price data skewness: {stats.skew(df_merged['adj_close'])}\")\n",
        "print(f\"Stock sentiment data skewness: {stats.skew(df_merged['avg_sentiment'])}\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Stock normality statistic=121.918, p=0.000\n",
            "Sentiment normality statistic=19.070, p=0.000\n",
            "\n",
            "Stock price data looks non-Gaussian\n",
            "Stock sentiment data looks non-Gaussian\n",
            "\n",
            "Stock price data skewness: -0.33355357511892075\n",
            "Stock sentiment data skewness: 0.3159856974541475\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 379
        },
        "id": "eOU45II-mBjO",
        "outputId": "648fdf99-7d7d-432f-89b1-053689f24c9f"
      },
      "source": [
        "# Plot correlation between adjusted close price and average sentiment\n",
        "corr = df_merged.corr(method='pearson')\n",
        "sns.heatmap(corr, xticklabels=corr.columns.values, yticklabels=corr.columns.values, annot=True, annot_kws={'size': 12})\n",
        "heat_map = plt.gcf()\n",
        "heat_map.set_size_inches(10,6)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x432 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uv3BbPSPtvIj",
        "outputId": "9bfea925-cd21-4ba7-d0ab-ed1bfcfda3f4"
      },
      "source": [
        "# Split merged dataset into train and test data\n",
        "n_obs=15\n",
        "X_train, X_test = df_merged[0:-n_obs], df_merged[-n_obs:]\n",
        "print(X_train.shape, X_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(629, 2) (15, 2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 547
        },
        "id": "EIUdbYFzqAzQ",
        "outputId": "990bcc7f-7866-41dc-cc2c-57de7e89d659"
      },
      "source": [
        "# Apply first differencing\n",
        "X_train_transformed = X_train.diff().dropna()\n",
        "print(X_train_transformed)\n",
        "X_train_transformed.describe()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "            adj_close  avg_sentiment\n",
            "date                                \n",
            "2017-03-23  -0.140228         0.5106\n",
            "2017-03-28   0.271698        -0.9510\n",
            "2017-03-29   0.131462         1.2720\n",
            "2017-03-30   0.078903         0.0000\n",
            "2017-03-31  -0.306778        -0.6209\n",
            "...               ...            ...\n",
            "2019-12-02   0.266205        -0.5894\n",
            "2019-12-03  -0.798599        -0.2139\n",
            "2019-12-04   0.275696         0.7045\n",
            "2019-12-05  -0.494362        -0.6386\n",
            "2019-12-06   0.836632         0.7104\n",
            "\n",
            "[628 rows x 2 columns]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "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>adj_close</th>\n",
              "      <th>avg_sentiment</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>628.000000</td>\n",
              "      <td>628.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>0.004584</td>\n",
              "      <td>0.000896</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>1.079783</td>\n",
              "      <td>0.295013</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>-5.242920</td>\n",
              "      <td>-0.951000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>-0.539167</td>\n",
              "      <td>-0.170050</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>0.054392</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>0.604785</td>\n",
              "      <td>0.154975</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>3.870285</td>\n",
              "      <td>1.272000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        adj_close  avg_sentiment\n",
              "count  628.000000     628.000000\n",
              "mean     0.004584       0.000896\n",
              "std      1.079783       0.295013\n",
              "min     -5.242920      -0.951000\n",
              "25%     -0.539167      -0.170050\n",
              "50%      0.054392       0.000000\n",
              "75%      0.604785       0.154975\n",
              "max      3.870285       1.272000"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 128
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "i69F0Vg9qpqM",
        "outputId": "896d9198-90ef-45ea-f478-28b417f4c464"
      },
      "source": [
        "# Check for stationarity (low p-value suggests no unit root -> stationary time series)\n",
        "# Code from https://towardsdatascience.com/granger-causality-and-vector-auto-regressive-model-for-time-series-forecasting-3226a64889a6\n",
        "from statsmodels.tsa.stattools import adfuller\n",
        "\n",
        "def augmented_dickey_fuller_statistics(time_series, name):\n",
        "  result = adfuller(time_series.values)\n",
        "  print(f\"Augmented Dickey-Fuller Test: {name} Time Series\")\n",
        "  print('ADF Statistic: %f' % result[0])\n",
        "  print('p-value: %f' % result[1])\n",
        "  print('Critical Values:')\n",
        "  for key, value in result[4].items():\n",
        "    print('\\t%s: %.3f' % (key, value))\n",
        "  \n",
        "augmented_dickey_fuller_statistics(X_train_transformed['adj_close'], 'Adjusted Close')\n",
        "augmented_dickey_fuller_statistics(X_train_transformed['avg_sentiment'], 'Average Sentiment')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Augmented Dickey-Fuller Test: Adjusted Close Time Series\n",
            "ADF Statistic: -22.435993\n",
            "p-value: 0.000000\n",
            "Critical Values:\n",
            "\t1%: -3.441\n",
            "\t5%: -2.866\n",
            "\t10%: -2.569\n",
            "Augmented Dickey-Fuller Test: Average Sentiment Time Series\n",
            "ADF Statistic: -10.556430\n",
            "p-value: 0.000000\n",
            "Critical Values:\n",
            "\t1%: -3.441\n",
            "\t5%: -2.866\n",
            "\t10%: -2.569\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 407
        },
        "id": "GhypnD5M04Tq",
        "outputId": "9071a8a3-8a22-4784-eed1-ca68171ada35"
      },
      "source": [
        "# Plot variation over time to check for stationarity\n",
        "# Code from https://towardsdatascience.com/granger-causality-and-vector-auto-regressive-model-for-time-series-forecasting-3226a64889a6\n",
        "fig, axes = plt.subplots(nrows=2, ncols=1, dpi=100)\n",
        "for i, ax in enumerate(axes.flatten()):\n",
        "  d = X_train_transformed[X_train_transformed.columns[i]]\n",
        "  ax.plot(d, color='red', linewidth=1)\n",
        "\n",
        "  # Decorations\n",
        "  ax.set_title(X_train_transformed.columns[i])\n",
        "  ax.xaxis.set_ticks_position('none')\n",
        "  ax.yaxis.set_ticks_position('none')\n",
        "  ax.spines['top'].set_alpha(0)\n",
        "  ax.tick_params(labelsize=6)\n",
        "plt.tight_layout();"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 600x400 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 127
        },
        "id": "LVVjFW75wL52",
        "outputId": "ca941570-6e08-4e65-e819-92d0bc5ec568"
      },
      "source": [
        "# Perform Granger causality test (p-value < 0.05 suggests x Granger causes y)\n",
        "# Code from https://stackoverflow.com/questions/58005681/is-it-possible-to-run-a-vector-autoregression-analysis-on-a-large-gdp-data-with\n",
        "from statsmodels.tsa.stattools import grangercausalitytests\n",
        "\n",
        "maxlag=12\n",
        "test = 'ssr-chi2test'\n",
        "\n",
        "def grangers_causality_matrix(data, variables, test = 'ssr_chi2test', verbose=False):\n",
        "\n",
        "    dataset = pd.DataFrame(np.zeros((len(variables), len(variables))), columns=variables, index=variables)\n",
        "    for c in dataset.columns:\n",
        "        for r in dataset.index:\n",
        "            test_result = grangercausalitytests(data[[r,c]], maxlag=maxlag, verbose=False)\n",
        "            p_values = [round(test_result[i+1][0][test][1],4) for i in range(maxlag)]\n",
        "            if verbose: print(f'Y = {r}, X = {c}, P Values = {p_values}')\n",
        "\n",
        "            min_p_value = np.min(p_values)\n",
        "            dataset.loc[r,c] = min_p_value\n",
        "\n",
        "    dataset.columns = [var+'_x' for var in variables]\n",
        "    dataset.index = [var+'_y' for var in variables]\n",
        "    return dataset\n",
        "\n",
        "print(X_train_transformed.columns)\n",
        "grangers_causality_matrix(X_train, variables = X_train.columns) "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Index(['adj_close', 'avg_sentiment'], dtype='object')\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "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>adj_close_x</th>\n",
              "      <th>avg_sentiment_x</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>adj_close_y</th>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.0313</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>avg_sentiment_y</th>\n",
              "      <td>0.0047</td>\n",
              "      <td>1.0000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                 adj_close_x  avg_sentiment_x\n",
              "adj_close_y           1.0000           0.0313\n",
              "avg_sentiment_y       0.0047           1.0000"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 131
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "opWJqzJJ19Ak",
        "outputId": "e81ae2da-8dc1-4ba9-a0fb-aff2669093b8"
      },
      "source": [
        "# Fit VAR model on train dataset\n",
        "# Code from https://towardsdatascience.com/granger-causality-and-vector-auto-regressive-model-for-time-series-forecasting-3226a64889a6\n",
        "from statsmodels.tsa.api import VAR\n",
        "\n",
        "mod = VAR(X_train_transformed)\n",
        "res = mod.fit(maxlags=15, ic='aic')\n",
        "print(res.summary())"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "  Summary of Regression Results   \n",
            "==================================\n",
            "Model:                         VAR\n",
            "Method:                        OLS\n",
            "Date:           Fri, 06, Aug, 2021\n",
            "Time:                     21:54:25\n",
            "--------------------------------------------------------------------\n",
            "No. of Equations:         2.00000    BIC:                   -2.44220\n",
            "Nobs:                     617.000    HQIC:                  -2.64383\n",
            "Log likelihood:          -849.779    FPE:                  0.0625352\n",
            "AIC:                     -2.77209    Det(Omega_mle):       0.0581213\n",
            "--------------------------------------------------------------------\n",
            "Results for equation adj_close\n",
            "====================================================================================\n",
            "                       coefficient       std. error           t-stat            prob\n",
            "------------------------------------------------------------------------------------\n",
            "const                     0.004357         0.043126            0.101           0.920\n",
            "L1.adj_close              0.115722         0.041042            2.820           0.005\n",
            "L1.avg_sentiment         -0.108192         0.194852           -0.555           0.579\n",
            "L2.adj_close              0.019014         0.041156            0.462           0.644\n",
            "L2.avg_sentiment         -0.390453         0.249392           -1.566           0.117\n",
            "L3.adj_close              0.045492         0.040811            1.115           0.265\n",
            "L3.avg_sentiment         -0.640847         0.282277           -2.270           0.023\n",
            "L4.adj_close             -0.032146         0.040784           -0.788           0.431\n",
            "L4.avg_sentiment         -0.299333         0.298489           -1.003           0.316\n",
            "L5.adj_close             -0.067235         0.040711           -1.652           0.099\n",
            "L5.avg_sentiment         -0.179829         0.309320           -0.581           0.561\n",
            "L6.adj_close             -0.048808         0.040760           -1.197           0.231\n",
            "L6.avg_sentiment         -0.097028         0.312752           -0.310           0.756\n",
            "L7.adj_close              0.051492         0.040820            1.261           0.207\n",
            "L7.avg_sentiment          0.035071         0.307102            0.114           0.909\n",
            "L8.adj_close              0.015752         0.040931            0.385           0.700\n",
            "L8.avg_sentiment         -0.157992         0.295390           -0.535           0.593\n",
            "L9.adj_close             -0.079140         0.040927           -1.934           0.053\n",
            "L9.avg_sentiment          0.294318         0.278814            1.056           0.291\n",
            "L10.adj_close            -0.048467         0.041015           -1.182           0.237\n",
            "L10.avg_sentiment        -0.210733         0.246501           -0.855           0.393\n",
            "L11.adj_close             0.019558         0.040813            0.479           0.632\n",
            "L11.avg_sentiment         0.219356         0.195161            1.124           0.261\n",
            "====================================================================================\n",
            "\n",
            "Results for equation avg_sentiment\n",
            "====================================================================================\n",
            "                       coefficient       std. error           t-stat            prob\n",
            "------------------------------------------------------------------------------------\n",
            "const                    -0.000080         0.009079           -0.009           0.993\n",
            "L1.adj_close              0.002956         0.008641            0.342           0.732\n",
            "L1.avg_sentiment         -0.811129         0.041023          -19.773           0.000\n",
            "L2.adj_close             -0.002208         0.008665           -0.255           0.799\n",
            "L2.avg_sentiment         -0.713281         0.052505          -13.585           0.000\n",
            "L3.adj_close              0.000520         0.008592            0.061           0.952\n",
            "L3.avg_sentiment         -0.592828         0.059429           -9.975           0.000\n",
            "L4.adj_close              0.002161         0.008586            0.252           0.801\n",
            "L4.avg_sentiment         -0.540015         0.062842           -8.593           0.000\n",
            "L5.adj_close             -0.008657         0.008571           -1.010           0.312\n",
            "L5.avg_sentiment         -0.427065         0.065122           -6.558           0.000\n",
            "L6.adj_close             -0.014147         0.008581           -1.649           0.099\n",
            "L6.avg_sentiment         -0.348963         0.065845           -5.300           0.000\n",
            "L7.adj_close             -0.005858         0.008594           -0.682           0.495\n",
            "L7.avg_sentiment         -0.294853         0.064655           -4.560           0.000\n",
            "L8.adj_close              0.009836         0.008617            1.141           0.254\n",
            "L8.avg_sentiment         -0.307032         0.062189           -4.937           0.000\n",
            "L9.adj_close             -0.001559         0.008617           -0.181           0.856\n",
            "L9.avg_sentiment         -0.221095         0.058699           -3.767           0.000\n",
            "L10.adj_close             0.004710         0.008635            0.545           0.585\n",
            "L10.avg_sentiment        -0.122513         0.051896           -2.361           0.018\n",
            "L11.adj_close             0.008934         0.008592            1.040           0.298\n",
            "L11.avg_sentiment        -0.090195         0.041088           -2.195           0.028\n",
            "====================================================================================\n",
            "\n",
            "Correlation matrix of residuals\n",
            "                 adj_close  avg_sentiment\n",
            "adj_close         1.000000       0.056689\n",
            "avg_sentiment     0.056689       1.000000\n",
            "\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/statsmodels/tsa/base/tsa_model.py:219: ValueWarning: A date index has been provided, but it has no associated frequency information and so will be ignored when e.g. forecasting.\n",
            "  ' ignored when e.g. forecasting.', ValueWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZzJ5U2bE3Z0I",
        "outputId": "3849c6b8-4e0f-4a65-aab5-47369d843105"
      },
      "source": [
        "# Calculate Durbin-Watson Statistic (range of 0 to 4, 1.5 to 2.5 suggests minimal auto-correlation)\n",
        "from statsmodels.stats.stattools import durbin_watson\n",
        "out = durbin_watson(res.resid)\n",
        "\n",
        "for col, val in zip(X_train_transformed.columns, out):\n",
        "  print((col), ':', round(val, 2))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "adj_close : 2.0\n",
            "avg_sentiment : 2.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 743
        },
        "id": "YimC8xOT390w",
        "outputId": "ddbb234b-d197-4a98-fb6b-df0cb3d10109"
      },
      "source": [
        "# Forecast future values using past lags\n",
        "lag_order = res.k_ar\n",
        "print(lag_order)\n",
        "\n",
        "# Input data for forecasting\n",
        "input_data = X_train_transformed.values[-lag_order:]\n",
        "print(input_data)\n",
        "\n",
        "# forecasting\n",
        "pred = res.forecast(y=input_data, steps=n_obs)\n",
        "pred = pd.DataFrame(pred, index=X_test.index, columns=X_test.columns + '_pred')\n",
        "pred"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "11\n",
            "[[-0.40881348  0.2028    ]\n",
            " [ 0.01903534  0.4092    ]\n",
            " [ 1.55918121 -0.6782    ]\n",
            " [-1.55918121  0.5478    ]\n",
            " [ 0.09506226 -0.2532    ]\n",
            " [ 0.21865845  0.5894    ]\n",
            " [ 0.26620483 -0.5894    ]\n",
            " [-0.79859924 -0.2139    ]\n",
            " [ 0.2756958   0.7045    ]\n",
            " [-0.49436188 -0.6386    ]\n",
            " [ 0.83663177  0.7104    ]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "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>adj_close_pred</th>\n",
              "      <th>avg_sentiment_pred</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>date</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2019-12-09</th>\n",
              "      <td>-0.553503</td>\n",
              "      <td>-0.414223</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-10</th>\n",
              "      <td>0.542553</td>\n",
              "      <td>-0.037020</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-11</th>\n",
              "      <td>-0.410116</td>\n",
              "      <td>0.083884</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-12</th>\n",
              "      <td>0.382869</td>\n",
              "      <td>-0.064605</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-13</th>\n",
              "      <td>-0.215524</td>\n",
              "      <td>0.089345</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-16</th>\n",
              "      <td>0.015724</td>\n",
              "      <td>-0.086389</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-17</th>\n",
              "      <td>0.367148</td>\n",
              "      <td>0.061229</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-18</th>\n",
              "      <td>-0.511458</td>\n",
              "      <td>-0.046869</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-19</th>\n",
              "      <td>0.512030</td>\n",
              "      <td>-0.018814</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-20</th>\n",
              "      <td>-0.358849</td>\n",
              "      <td>0.099577</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-23</th>\n",
              "      <td>0.192319</td>\n",
              "      <td>-0.079849</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-24</th>\n",
              "      <td>-0.027475</td>\n",
              "      <td>0.051364</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-26</th>\n",
              "      <td>-0.105912</td>\n",
              "      <td>-0.029871</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-30</th>\n",
              "      <td>0.098611</td>\n",
              "      <td>-0.007114</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2019-12-31</th>\n",
              "      <td>-0.080811</td>\n",
              "      <td>0.018485</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            adj_close_pred  avg_sentiment_pred\n",
              "date                                          \n",
              "2019-12-09       -0.553503           -0.414223\n",
              "2019-12-10        0.542553           -0.037020\n",
              "2019-12-11       -0.410116            0.083884\n",
              "2019-12-12        0.382869           -0.064605\n",
              "2019-12-13       -0.215524            0.089345\n",
              "2019-12-16        0.015724           -0.086389\n",
              "2019-12-17        0.367148            0.061229\n",
              "2019-12-18       -0.511458           -0.046869\n",
              "2019-12-19        0.512030           -0.018814\n",
              "2019-12-20       -0.358849            0.099577\n",
              "2019-12-23        0.192319           -0.079849\n",
              "2019-12-24       -0.027475            0.051364\n",
              "2019-12-26       -0.105912           -0.029871\n",
              "2019-12-30        0.098611           -0.007114\n",
              "2019-12-31       -0.080811            0.018485"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 134
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Llh0UGfe4tFE"
      },
      "source": [
        "# De-difference forecasts from the training data scale\n",
        "def invert_transformation(X_train, pred):\n",
        "  forecast = pred.copy()\n",
        "  columns = X_train.columns\n",
        "  for col in columns:\n",
        "    forecast[str(col)+'_pred'] = X_train[col].iloc[-1] + forecast[str(col)+'_pred'].cumsum()\n",
        "  return forecast\n",
        "  \n",
        "output = invert_transformation(X_train, pred)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 511
        },
        "id": "-a4aCybl5BZf",
        "outputId": "440a2bab-0a4d-4497-82d7-f1521dab424c"
      },
      "source": [
        "# Merge predicted and true data set\n",
        "merged_forecasts = pd.concat([output['adj_close_pred'], X_test['adj_close']], axis=1)\n",
        "merged_forecasts['accuracy'] = round(merged_forecasts.apply(lambda row: row['adj_close_pred'] / row['adj_close'] * 100, axis = 1),2)\n",
        "merged_forecasts['accuracy'] = pd.Series([\"{0:.2f}%\".format(val) for val in merged_forecasts['accuracy']], index = merged_forecasts.index)\n",
        "merged_forecasts = merged_forecasts.round(decimals=2)\n",
        "merged_forecasts = merged_forecasts.reset_index()\n",
        "merged_forecasts = merged_forecasts.sort_values(by='date', ascending=False)\n",
        "merged_forecasts"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>adj_close_pred</th>\n",
              "      <th>adj_close</th>\n",
              "      <th>accuracy</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>2019-12-31</td>\n",
              "      <td>42.42</td>\n",
              "      <td>43.40</td>\n",
              "      <td>97.74%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>2019-12-30</td>\n",
              "      <td>42.45</td>\n",
              "      <td>43.60</td>\n",
              "      <td>97.36%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>2019-12-27</td>\n",
              "      <td>42.41</td>\n",
              "      <td>43.52</td>\n",
              "      <td>97.44%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>2019-12-26</td>\n",
              "      <td>42.41</td>\n",
              "      <td>43.40</td>\n",
              "      <td>97.73%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>2019-12-24</td>\n",
              "      <td>42.45</td>\n",
              "      <td>42.97</td>\n",
              "      <td>98.77%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>2019-12-23</td>\n",
              "      <td>42.37</td>\n",
              "      <td>42.56</td>\n",
              "      <td>99.55%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>2019-12-20</td>\n",
              "      <td>42.30</td>\n",
              "      <td>42.55</td>\n",
              "      <td>99.43%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>2019-12-19</td>\n",
              "      <td>42.33</td>\n",
              "      <td>42.11</td>\n",
              "      <td>100.51%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>2019-12-18</td>\n",
              "      <td>42.24</td>\n",
              "      <td>41.68</td>\n",
              "      <td>101.35%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>2019-12-17</td>\n",
              "      <td>42.31</td>\n",
              "      <td>41.87</td>\n",
              "      <td>101.04%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2019-12-16</td>\n",
              "      <td>42.37</td>\n",
              "      <td>42.01</td>\n",
              "      <td>100.86%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2019-12-13</td>\n",
              "      <td>42.19</td>\n",
              "      <td>42.06</td>\n",
              "      <td>100.32%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2019-12-12</td>\n",
              "      <td>42.22</td>\n",
              "      <td>41.51</td>\n",
              "      <td>101.73%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2019-12-11</td>\n",
              "      <td>42.44</td>\n",
              "      <td>40.60</td>\n",
              "      <td>104.53%</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2019-12-10</td>\n",
              "      <td>42.36</td>\n",
              "      <td>41.27</td>\n",
              "      <td>102.63%</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "          date  adj_close_pred  adj_close accuracy\n",
              "14  2019-12-31           42.42      43.40   97.74%\n",
              "13  2019-12-30           42.45      43.60   97.36%\n",
              "12  2019-12-27           42.41      43.52   97.44%\n",
              "11  2019-12-26           42.41      43.40   97.73%\n",
              "10  2019-12-24           42.45      42.97   98.77%\n",
              "9   2019-12-23           42.37      42.56   99.55%\n",
              "8   2019-12-20           42.30      42.55   99.43%\n",
              "7   2019-12-19           42.33      42.11  100.51%\n",
              "6   2019-12-18           42.24      41.68  101.35%\n",
              "5   2019-12-17           42.31      41.87  101.04%\n",
              "4   2019-12-16           42.37      42.01  100.86%\n",
              "3   2019-12-13           42.19      42.06  100.32%\n",
              "2   2019-12-12           42.22      41.51  101.73%\n",
              "1   2019-12-11           42.44      40.60  104.53%\n",
              "0   2019-12-10           42.36      41.27  102.63%"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 120
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kNPDDLyT_WCx",
        "outputId": "ec9ab13c-97c8-4d02-af12-5dc5df754416"
      },
      "source": [
        "# Compute bias and errors\n",
        "from sklearn.metrics import mean_absolute_error as mae, mean_squared_error as mse\n",
        "\n",
        "forecast_errors = [merged_forecasts['adj_close'][i]- merged_forecasts['adj_close_pred'][i] for i in range(len(merged_forecasts['adj_close']))]\n",
        "bias = sum(forecast_errors) * 1.0/len(merged_forecasts['adj_close'])\n",
        "\n",
        "print('Bias: %f' % bias)\n",
        "print('Mean absolute error:', mae(merged_forecasts['adj_close'].values, merged_forecasts['adj_close_pred'].values))\n",
        "print('Mean squared error:', mse(merged_forecasts['adj_close'].values, merged_forecasts['adj_close_pred'].values))\n",
        "print('Root mean squared error:', math.sqrt(mse(merged_forecasts['adj_close'].values, merged_forecasts['adj_close_pred'].values)))\n",
        "# Good predictions at num_tweets=1000000: T (MAE=0.301), CMCSA (MAE=0.67), CVS (MAE=0.901)\n",
        "# Good predictions at num_tweets=2000000: EBAY (MAE=0.775), CVS (MAE=0.703)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Bias: -0.010667\n",
            "Mean absolute error: 0.7026666666666667\n",
            "Mean squared error: 0.7095999999999995\n",
            "Root mean squared error: 0.8423775875461071\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}