ML_Project_Time_Series(V2).ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/hogan-tech/123-Essential-JavaScript-Interview-Questions/blob/master/ML_Project_Time_Series(V2).ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Import data"
      ],
      "metadata": {
        "id": "aU0l0MiFQJSq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install kaggle"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "J06XMeCWnOdq",
        "outputId": "223806f9-7584-42c4-b18d-0cda17419431"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: kaggle in /usr/local/lib/python3.10/dist-packages (1.6.17)\n",
            "Requirement already satisfied: six>=1.10 in /usr/local/lib/python3.10/dist-packages (from kaggle) (1.16.0)\n",
            "Requirement already satisfied: certifi>=2023.7.22 in /usr/local/lib/python3.10/dist-packages (from kaggle) (2024.8.30)\n",
            "Requirement already satisfied: python-dateutil in /usr/local/lib/python3.10/dist-packages (from kaggle) (2.8.2)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from kaggle) (2.32.3)\n",
            "Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (from kaggle) (4.66.6)\n",
            "Requirement already satisfied: python-slugify in /usr/local/lib/python3.10/dist-packages (from kaggle) (8.0.4)\n",
            "Requirement already satisfied: urllib3 in /usr/local/lib/python3.10/dist-packages (from kaggle) (2.2.3)\n",
            "Requirement already satisfied: bleach in /usr/local/lib/python3.10/dist-packages (from kaggle) (6.2.0)\n",
            "Requirement already satisfied: webencodings in /usr/local/lib/python3.10/dist-packages (from bleach->kaggle) (0.5.1)\n",
            "Requirement already satisfied: text-unidecode>=1.3 in /usr/local/lib/python3.10/dist-packages (from python-slugify->kaggle) (1.3)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->kaggle) (3.4.0)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->kaggle) (3.10)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install --upgrade holidays"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KA9NKwaXsP-O",
        "outputId": "8b3c8400-bf36-4969-fcb0-0f7e903c2653"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: holidays in /usr/local/lib/python3.10/dist-packages (0.60)\n",
            "Requirement already satisfied: python-dateutil in /usr/local/lib/python3.10/dist-packages (from holidays) (2.8.2)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil->holidays) (1.16.0)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WM84WJUBmHLR"
      },
      "outputs": [],
      "source": [
        "#Importing all the required libraries\n",
        "import tensorflow as tf\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import sklearn.preprocessing\n",
        "import holidays\n",
        "import os\n",
        "import seaborn as sns"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Import Hourly Energy Consumption Dataset\n",
        "\n",
        "The hourly energy consumption datasets consist of energy (megawatt units) of different power companies in the US."
      ],
      "metadata": {
        "id": "qN34xj6WQWkL"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Loading the dataset\n",
        "import kagglehub\n",
        "\n",
        "path = kagglehub.dataset_download(\"robikscube/hourly-energy-consumption\")\n",
        "\n",
        "print(\"Path to dataset files:\", path)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ulqv8UdvnP1P",
        "outputId": "37ed5170-510c-4a19-9ab1-3ca192b52853"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading from https://www.kaggle.com/api/v1/datasets/download/robikscube/hourly-energy-consumption?dataset_version_number=3...\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 11.4M/11.4M [00:00<00:00, 37.3MB/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting files...\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Path to dataset files: /root/.cache/kagglehub/datasets/robikscube/hourly-energy-consumption/versions/3\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Import holiday dataset\n",
        "We want to use external US holiday dataset to see if we can find a pattern of energy consumption around holiday days"
      ],
      "metadata": {
        "id": "mXRa-hoGQMlt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "path_holiday = kagglehub.dataset_download(\"donnetew/us-holiday-dates-2004-2021\")"
      ],
      "metadata": {
        "id": "2e3smK2woO1y",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e6fbd611-5ef2-4fc3-8e76-dd8dad243ed1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading from https://www.kaggle.com/api/v1/datasets/download/donnetew/us-holiday-dates-2004-2021?dataset_version_number=1...\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|██████████| 2.45k/2.45k [00:00<00:00, 4.23MB/s]"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Extracting files...\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Path to holiday dataset:\", path_holiday)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8grlkb53ohpM",
        "outputId": "5cff5cdd-f191-4e77-8912-391bdd57293f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Path to holiday dataset: /root/.cache/kagglehub/datasets/donnetew/us-holiday-dates-2004-2021/versions/1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!ls /root/.cache/kagglehub/datasets/robikscube/hourly-energy-consumption/versions/3\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eCaPkzfynxkn",
        "outputId": "56888620-cdc5-4ef4-fe4f-b7a0a008f55e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "AEP_hourly.csv\t   DEOK_hourly.csv  EKPC_hourly.csv\tNI_hourly.csv\t    PJM_Load_hourly.csv\n",
            "COMED_hourly.csv   DOM_hourly.csv   est_hourly.paruqet\tPJME_hourly.csv     PJMW_hourly.csv\n",
            "DAYTON_hourly.csv  DUQ_hourly.csv   FE_hourly.csv\tpjm_hourly_est.csv\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 1. Data Exploration and Analysis\n"
      ],
      "metadata": {
        "id": "QIhDSnvppNAr"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.1 Merging the energy consumption datasets\n",
        "By Yian"
      ],
      "metadata": {
        "id": "FAJIUlcvQ62W"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "aep = pd.read_csv(os.path.join(path, \"AEP_hourly.csv\"))\n",
        "comed = pd.read_csv(os.path.join(path, \"COMED_hourly.csv\"))\n",
        "dayton = pd.read_csv(os.path.join(path, \"DAYTON_hourly.csv\"))\n",
        "deok = pd.read_csv(os.path.join(path, \"DEOK_hourly.csv\"))\n",
        "dom = pd.read_csv(os.path.join(path, \"DOM_hourly.csv\"))\n",
        "duq = pd.read_csv(os.path.join(path, \"DUQ_hourly.csv\"))\n",
        "ekpc = pd.read_csv(os.path.join(path, \"EKPC_hourly.csv\"))\n",
        "fe = pd.read_csv(os.path.join(path, \"FE_hourly.csv\"))\n",
        "ni = pd.read_csv(os.path.join(path, \"NI_hourly.csv\"))\n",
        "pjme = pd.read_csv(os.path.join(path, \"PJME_hourly.csv\"))\n",
        "pjm = pd.read_csv(os.path.join(path, \"pjm_hourly_est.csv\")) # diff format\n",
        "pjm_load = pd.read_csv(os.path.join(path, \"PJM_Load_hourly.csv\"))\n",
        "pjmw = pd.read_csv(os.path.join(path, \"PJMW_hourly.csv\"))"
      ],
      "metadata": {
        "id": "7B6beNp1scOm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "aep[\"company\"] = \"AEP_MW\"\n",
        "comed[\"company\"] = \"COMED_MW\"\n",
        "dayton[\"company\"] = \"DAYTON_MW\"\n",
        "deok[\"company\"] = \"DEOK_MW\"\n",
        "dom[\"company\"] = \"DOM_MW\"\n",
        "duq[\"company\"] = \"DUQ_MW\"\n",
        "ekpc[\"company\"] = \"EKPC_MW\"\n",
        "fe[\"company\"] = \"FE_MW\"\n",
        "ni[\"company\"] = \"NI_MW\"\n",
        "pjme[\"company\"] = \"PJME_MW\"\n",
        "pjmw[\"company\"] = \"PJMW_MW\"\n",
        "pjm_load[\"company\"] = \"PJM_Load_MW\""
      ],
      "metadata": {
        "id": "fm2xP5i6zWT8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "aep = aep.rename(columns={\"AEP_MW\": \"MW\"})\n",
        "comed = comed.rename(columns={\"COMED_MW\": \"MW\"})\n",
        "dayton = dayton.rename(columns={\"DAYTON_MW\": \"MW\"})\n",
        "deok = deok.rename(columns={\"DEOK_MW\": \"MW\"})\n",
        "dom = dom.rename(columns={\"DOM_MW\": \"MW\"})\n",
        "duq = duq.rename(columns={\"DUQ_MW\": \"MW\"})\n",
        "ekpc = ekpc.rename(columns={\"EKPC_MW\": \"MW\"})\n",
        "fe = fe.rename(columns={\"FE_MW\": \"MW\"})\n",
        "ni = ni.rename(columns={\"NI_MW\": \"MW\"})\n",
        "pjme = pjme.rename(columns={\"PJME_MW\": \"MW\"})\n",
        "pjmw = pjmw.rename(columns={\"PJMW_MW\": \"MW\"})\n",
        "pjm_load = pjm_load.rename(columns={\"PJM_Load_MW\": \"MW\"})"
      ],
      "metadata": {
        "id": "QUcPK4BNwIrz"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_combined = pd.concat([aep, comed, dayton, deok, dom, duq, ekpc, fe, ni, pjme, pjmw, pjm_load], ignore_index=True)\n",
        "df_combined.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "QFPG81ALwlaU",
        "outputId": "510514fa-1069-4e5b-8d1e-f69537cdb1fe"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "              Datetime       MW company\n",
              "0  2004-12-31 01:00:00  13478.0  AEP_MW\n",
              "1  2004-12-31 02:00:00  12865.0  AEP_MW\n",
              "2  2004-12-31 03:00:00  12577.0  AEP_MW\n",
              "3  2004-12-31 04:00:00  12517.0  AEP_MW\n",
              "4  2004-12-31 05:00:00  12670.0  AEP_MW"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-8c5d7c78-0d14-4f80-afde-492eb0e0452f\" class=\"colab-df-container\">\n",
              "    <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>Datetime</th>\n",
              "      <th>MW</th>\n",
              "      <th>company</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2004-12-31 01:00:00</td>\n",
              "      <td>13478.0</td>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2004-12-31 02:00:00</td>\n",
              "      <td>12865.0</td>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2004-12-31 03:00:00</td>\n",
              "      <td>12577.0</td>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2004-12-31 04:00:00</td>\n",
              "      <td>12517.0</td>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2004-12-31 05:00:00</td>\n",
              "      <td>12670.0</td>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-8c5d7c78-0d14-4f80-afde-492eb0e0452f')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-8c5d7c78-0d14-4f80-afde-492eb0e0452f button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-8c5d7c78-0d14-4f80-afde-492eb0e0452f');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-b7221889-4d0e-47d5-81af-4562a7b65888\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-b7221889-4d0e-47d5-81af-4562a7b65888')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-b7221889-4d0e-47d5-81af-4562a7b65888 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "df_combined"
            }
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_combined.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Nd47cNqY36mj",
        "outputId": "f8665df1-c58c-4806-e65c-1a64e65205e6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1090167, 3)"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_c = df_combined.copy()"
      ],
      "metadata": {
        "id": "K9jo-NDhIYlG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_c .columns"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1hTW5NyJIhqL",
        "outputId": "efff5d9e-05eb-4fae-fca0-467adc2014b2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Index(['Datetime', 'MW', 'company'], dtype='object')"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.2 Check dataset type and statistical breakdown\n",
        "By Yian"
      ],
      "metadata": {
        "id": "qn35k3VRsHtC"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_c .info()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fdW3rBLrIiej",
        "outputId": "f886bcf4-c201-4036-8cb7-1e9861a8b8cb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 1090167 entries, 0 to 1090166\n",
            "Data columns (total 3 columns):\n",
            " #   Column    Non-Null Count    Dtype  \n",
            "---  ------    --------------    -----  \n",
            " 0   Datetime  1090167 non-null  object \n",
            " 1   MW        1090167 non-null  float64\n",
            " 2   company   1090167 non-null  object \n",
            "dtypes: float64(1), object(2)\n",
            "memory usage: 25.0+ MB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_c .describe()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "QdLOTHQ9ItTd",
        "outputId": "d67572cd-0a25-4988-dd0b-59a08289e379"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                 MW\n",
              "count  1.090167e+06\n",
              "mean   1.120959e+04\n",
              "std    1.053440e+04\n",
              "min    0.000000e+00\n",
              "25%    2.455000e+03\n",
              "50%    8.217000e+03\n",
              "75%    1.466500e+04\n",
              "max    6.200900e+04"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f0a67660-b095-43f7-a162-cf2c492d845b\" class=\"colab-df-container\">\n",
              "    <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>MW</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>count</th>\n",
              "      <td>1.090167e+06</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mean</th>\n",
              "      <td>1.120959e+04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>std</th>\n",
              "      <td>1.053440e+04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>min</th>\n",
              "      <td>0.000000e+00</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>25%</th>\n",
              "      <td>2.455000e+03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50%</th>\n",
              "      <td>8.217000e+03</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>75%</th>\n",
              "      <td>1.466500e+04</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>max</th>\n",
              "      <td>6.200900e+04</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f0a67660-b095-43f7-a162-cf2c492d845b')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-f0a67660-b095-43f7-a162-cf2c492d845b button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-f0a67660-b095-43f7-a162-cf2c492d845b');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-4331cbbb-3844-427d-88dd-29f91abb71d4\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-4331cbbb-3844-427d-88dd-29f91abb71d4')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-4331cbbb-3844-427d-88dd-29f91abb71d4 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "summary": "{\n  \"name\": \"df_c \",\n  \"rows\": 8,\n  \"fields\": [\n    {\n      \"column\": \"MW\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 380424.06324977847,\n        \"min\": 0.0,\n        \"max\": 1090167.0,\n        \"num_unique_values\": 8,\n        \"samples\": [\n          11209.588650179285,\n          8217.0,\n          1090167.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Check if there are null values"
      ],
      "metadata": {
        "id": "uQb7SbwDsBnq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# no null values in dataset\n",
        "df_c .isna().sum()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 178
        },
        "id": "mqYxk9onKruK",
        "outputId": "35b2c600-ed3b-40eb-e459-42099547f439"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Datetime    0\n",
              "MW          0\n",
              "company     0\n",
              "dtype: int64"
            ],
            "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>0</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>Datetime</th>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>MW</th>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>company</th>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div><br><label><b>dtype:</b> int64</label>"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [],
      "metadata": {
        "id": "qMn-oIgqr_KC"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.4 Separate categorical from numerical attributes for data visualization\n",
        "By Yian"
      ],
      "metadata": {
        "id": "I8tnlWtXtIKW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "numerical_attributes = df_c .select_dtypes( include=['float64'] )\n",
        "numerical_attributes.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "eqa607yerCOg",
        "outputId": "0c8c2c2a-b688-458b-ba86-8beb2a1c7760"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        MW\n",
              "0  13478.0\n",
              "1  12865.0\n",
              "2  12577.0\n",
              "3  12517.0\n",
              "4  12670.0"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-61c54550-c0b6-49ce-ba1d-ccfcae4dba81\" class=\"colab-df-container\">\n",
              "    <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>MW</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>13478.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>12865.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>12577.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>12517.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>12670.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-61c54550-c0b6-49ce-ba1d-ccfcae4dba81')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-61c54550-c0b6-49ce-ba1d-ccfcae4dba81 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-61c54550-c0b6-49ce-ba1d-ccfcae4dba81');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-ee52a34e-6bd0-4954-ace3-25fe8e3a680d\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-ee52a34e-6bd0-4954-ace3-25fe8e3a680d')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-ee52a34e-6bd0-4954-ace3-25fe8e3a680d button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "numerical_attributes"
            }
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "categorical_attributes = df_c [\"company\"]\n",
        "categorical_attributes.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 241
        },
        "id": "UQ0wg1FlrQzw",
        "outputId": "00f5e375-3941-400e-e162-3e6a4d5255ad"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0    AEP_MW\n",
              "1    AEP_MW\n",
              "2    AEP_MW\n",
              "3    AEP_MW\n",
              "4    AEP_MW\n",
              "Name: company, dtype: object"
            ],
            "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>company</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>AEP_MW</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div><br><label><b>dtype:</b> object</label>"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.5 Data Analysis\n",
        "By Yian\n",
        "<br>\n",
        "Get the central tendency mean and median of the energy consumption dataset to check the typical data point. Check standard deviation, skewness and kurtosis of the dataset"
      ],
      "metadata": {
        "id": "KsM0k9g_vIDg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "num_att_mean = pd.DataFrame(numerical_attributes.apply(np.mean)).T\n",
        "num_att_median = pd.DataFrame(numerical_attributes.apply(np.median)).T\n",
        "d1 = pd.DataFrame(numerical_attributes.apply(np.std)).T # standard deviation\n",
        "d2 = pd.DataFrame(numerical_attributes.apply(min)).T\n",
        "d3 = pd.DataFrame(numerical_attributes.apply(max)).T\n",
        "d4 = pd.DataFrame(numerical_attributes.apply(lambda x: x.max() - x.min())).T # range\n",
        "d5 = pd.DataFrame(numerical_attributes.apply(lambda x: x.skew())).T # skewness\n",
        "d6 = pd.DataFrame(numerical_attributes.apply(lambda x: x.kurtosis())).T # kurtosis\n"
      ],
      "metadata": {
        "id": "TW7j3-9GsYuq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "stats_df = pd.concat([d1,d2,d3,d4,d5,d6,num_att_mean,num_att_median]).T"
      ],
      "metadata": {
        "id": "F_p8v_gRwfgG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# rename columns\n",
        "stats_df.columns = [\"std\", \"min\", \"max\", \"range\", \"skew\", \"kurtosis\", \"mean\", \"median\"]"
      ],
      "metadata": {
        "id": "ENxspXN2w0rZ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "stats_df"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        },
        "id": "6B5EJ3aEw_Al",
        "outputId": "d3c7cc8e-8176-4456-b2d8-63437f00a94b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "             std  min      max    range      skew  kurtosis         mean  \\\n",
              "MW  10534.394704  0.0  62009.0  62009.0  1.356423  1.198281  11209.58865   \n",
              "\n",
              "    median  \n",
              "MW  8217.0  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f346d8ec-32e1-4ca7-9208-b2486f530059\" class=\"colab-df-container\">\n",
              "    <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>std</th>\n",
              "      <th>min</th>\n",
              "      <th>max</th>\n",
              "      <th>range</th>\n",
              "      <th>skew</th>\n",
              "      <th>kurtosis</th>\n",
              "      <th>mean</th>\n",
              "      <th>median</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>MW</th>\n",
              "      <td>10534.394704</td>\n",
              "      <td>0.0</td>\n",
              "      <td>62009.0</td>\n",
              "      <td>62009.0</td>\n",
              "      <td>1.356423</td>\n",
              "      <td>1.198281</td>\n",
              "      <td>11209.58865</td>\n",
              "      <td>8217.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f346d8ec-32e1-4ca7-9208-b2486f530059')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-f346d8ec-32e1-4ca7-9208-b2486f530059 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-f346d8ec-32e1-4ca7-9208-b2486f530059');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "  <div id=\"id_0e407508-a64c-4fb4-83b2-d2b9697c04d9\">\n",
              "    <style>\n",
              "      .colab-df-generate {\n",
              "        background-color: #E8F0FE;\n",
              "        border: none;\n",
              "        border-radius: 50%;\n",
              "        cursor: pointer;\n",
              "        display: none;\n",
              "        fill: #1967D2;\n",
              "        height: 32px;\n",
              "        padding: 0 0 0 0;\n",
              "        width: 32px;\n",
              "      }\n",
              "\n",
              "      .colab-df-generate:hover {\n",
              "        background-color: #E2EBFA;\n",
              "        box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "        fill: #174EA6;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate {\n",
              "        background-color: #3B4455;\n",
              "        fill: #D2E3FC;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate:hover {\n",
              "        background-color: #434B5C;\n",
              "        box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "        filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "        fill: #FFFFFF;\n",
              "      }\n",
              "    </style>\n",
              "    <button class=\"colab-df-generate\" onclick=\"generateWithVariable('stats_df')\"\n",
              "            title=\"Generate code using this dataframe.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M7,19H8.4L18.45,9,17,7.55,7,17.6ZM5,21V16.75L18.45,3.32a2,2,0,0,1,2.83,0l1.4,1.43a1.91,1.91,0,0,1,.58,1.4,1.91,1.91,0,0,1-.58,1.4L9.25,21ZM18.45,9,17,7.55Zm-12,3A5.31,5.31,0,0,0,4.9,8.1,5.31,5.31,0,0,0,1,6.5,5.31,5.31,0,0,0,4.9,4.9,5.31,5.31,0,0,0,6.5,1,5.31,5.31,0,0,0,8.1,4.9,5.31,5.31,0,0,0,12,6.5,5.46,5.46,0,0,0,6.5,12Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "    <script>\n",
              "      (() => {\n",
              "      const buttonEl =\n",
              "        document.querySelector('#id_0e407508-a64c-4fb4-83b2-d2b9697c04d9 button.colab-df-generate');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      buttonEl.onclick = () => {\n",
              "        google.colab.notebook.generateWithVariable('stats_df');\n",
              "      }\n",
              "      })();\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "dataframe",
              "variable_name": "stats_df",
              "summary": "{\n  \"name\": \"stats_df\",\n  \"rows\": 1,\n  \"fields\": [\n    {\n      \"column\": \"std\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 10534.394703789007,\n        \"max\": 10534.394703789007,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          10534.394703789007\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"min\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 0.0,\n        \"max\": 0.0,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          0.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"max\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 62009.0,\n        \"max\": 62009.0,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          62009.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"range\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 62009.0,\n        \"max\": 62009.0,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          62009.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"skew\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 1.356423142819272,\n        \"max\": 1.356423142819272,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          1.356423142819272\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"kurtosis\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 1.1982814588872657,\n        \"max\": 1.1982814588872657,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          1.1982814588872657\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"mean\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 11209.588650179285,\n        \"max\": 11209.588650179285,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          11209.588650179285\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"median\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": null,\n        \"min\": 8217.0,\n        \"max\": 8217.0,\n        \"num_unique_values\": 1,\n        \"samples\": [\n          8217.0\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
            }
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.6 Check the distribution of the data\n",
        "By Yian"
      ],
      "metadata": {
        "id": "KWsdpF1-x7xz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import seaborn as sns\n",
        "sns.histplot( df_c['MW'] )\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        },
        "id": "Kw2Lg3OyxPmt",
        "outputId": "34c5d9c1-0c60-45fb-d09c-31353922a2c9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<Axes: xlabel='MW', ylabel='Count'>"
            ]
          },
          "metadata": {},
          "execution_count": 25
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.7 Data Analysis\n",
        "By Yian\n",
        "<br>\n",
        "<br>\n",
        "From the histogram we can see the distribution of the data, which shows a right-skewed distribution. The high standard deviation (10534.39) also indicates a large spread and variability in the energy consumption values.  Most of the data points are concentrated in the lower range, with a long tail extending to the right. The right tail indicates some extreme outliers, where energy consumption is high in some cases on certain dates, pulling the distribution to the right. We shall analysis what are the causes of these outliers and apply normalization technique to help our predictive modeling. Beause our data has a high range and standard deviation, it is better to normalize the data first so our model can converge more efficiently when learning the patterns of the training data. The mean of the dataset is 11209.25, which is higher than the median (8217.00), which is also typical in right-skewed distribution in this case."
      ],
      "metadata": {
        "id": "7YNe2ps7x92A"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(14, 8))\n",
        "sns.boxplot( x=\"company\", y=\"MW\", data=df_c, palette=\"pastel\", hue=\"company\")\n",
        "plt.xticks(rotation=45, ha=\"right\", fontsize=10)\n",
        "plt.title(\"MW Distribution by Company\")\n",
        "plt.xlabel(\"Electric Company\")\n",
        "plt.ylabel(\"Energy Consumption in MW\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 783
        },
        "id": "1Wbfq8Ne0b0Q",
        "outputId": "d399cf79-84a9-4f76-e950-7ae4d75b823a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1400x800 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.8 Check how the data varies on daily, weekly and monthly basis\n",
        "by pranay\n"
      ],
      "metadata": {
        "id": "8utjw1ug5Kr8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df1 = pjme.copy()\n",
        "df1 = df1.set_index('Datetime')\n",
        "df1.index = pd.to_datetime(df1.index)"
      ],
      "metadata": {
        "id": "d3LB37OP7Lp5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 176
        },
        "outputId": "81defd00-8c2e-4239-d380-f068acacff1d"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "name 'pjme' is not defined",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-2-c7685220e2a6>\u001b[0m in \u001b[0;36m<cell line: 1>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mdf1\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpjme\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mdf1\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdf1\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_index\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Datetime'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mdf1\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto_datetime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdf1\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'pjme' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import seaborn as sns\n",
        "df1['hour'] = df1.index.hour\n",
        "df1['dayofweek'] = df1.index.dayofweek\n",
        "df1['quarter'] = df1.index.quarter\n",
        "df1['month'] = df1.index.month\n",
        "df1['year'] = df1.index.year\n",
        "df1['dayofyear'] = df1.index.dayofyear\n",
        "df1['dayofmonth'] = df1.index.day\n",
        "df1['weekofyear'] = df1.index.isocalendar().week\n",
        "print(df1.head())\n",
        "\n",
        "def create_features(df):\n",
        "    df['Datetime'] = pd.to_datetime(df['Datetime'])\n",
        "    df.set_index('Datetime', inplace=True)\n",
        "    df = df.copy()\n",
        "    df['hour'] = df.index.hour\n",
        "    df['dayofweek'] = df.index.dayofweek\n",
        "    df['quarter'] = df.index.quarter\n",
        "    df['month'] = df.index.month\n",
        "    df['year'] = df.index.year\n",
        "    df['dayofyear'] = df.index.dayofyear\n",
        "    df['dayofmonth'] = df.index.day\n",
        "    df['weekofyear'] = df.index.isocalendar().week\n",
        "    return df\n",
        "\n",
        "# df = create_features(df)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 211
        },
        "id": "KqNi4twejoYC",
        "outputId": "088e75e0-e73e-49cb-ee2d-de29726e4f52"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "name 'df1' is not defined",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-1-22b32e66a6dd>\u001b[0m in \u001b[0;36m<cell line: 2>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mseaborn\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0msns\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mdf1\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'hour'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdf1\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhour\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mdf1\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'dayofweek'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdf1\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdayofweek\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mdf1\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'quarter'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdf1\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mquarter\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mdf1\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'month'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdf1\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmonth\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'df1' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Hourly analysis\n",
        "fig, ax = plt.subplots(figsize=(10, 8))\n",
        "sns.boxplot(data=df1, x='hour', y='MW')\n",
        "ax.set_title('MW by Hour')\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "I7RfMe61kBna"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#weekly analysis\n",
        "fig, ax = plt.subplots(figsize=(15, 10))\n",
        "sns.boxplot(data=df1, x='weekofyear', y='MW')\n",
        "ax.set_title('MW by week')\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "HRfvDRSrkHM-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Montly analysis\n",
        "fig, ax = plt.subplots(figsize=(10, 8))\n",
        "sns.boxplot(data=df1, x='month', y='MW', palette='Blues',legend=False,hue='month')\n",
        "ax.set_title('MW by Month')\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "m7Gs_yCIkJc-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 1.8.1 Analysis report\n",
        "by pranay\n",
        "\n",
        "When we check these boxplots , we can come to a conclusion that\n",
        "- hourly report says that during the mornings the usage goes up and stays up until night when it goes low so we can conclude that the day time power usage is more than the night time usage.\n",
        "-Monthy report says that during the mid months june,july and august, the power consumption is high while april and may has low power consumption.\n",
        "- We can take some insights and continue our data analysis further ."
      ],
      "metadata": {
        "id": "cS3WV2uRgzOd"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "0cGNPB4rzhpH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df"
      ],
      "metadata": {
        "id": "43_ifHA-06nD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "train = df1.loc[df1.index < '01-01-2015']\n",
        "test = df1.loc[df1.index >= '01-01-2015']\n",
        "\n",
        "fig, ax = plt.subplots(figsize=(15, 5))\n",
        "train.plot(ax=ax, label='Training Set', title='Data Train/Test Split')\n",
        "test.plot(ax=ax, label='Test Set')\n",
        "ax.axvline('01-01-2015', color='black', ls='--')\n",
        "ax.legend(['Training Set', 'Test Set'])\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "DGNLgdKBogSd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train = create_features(train)\n",
        "test = create_features(test)\n",
        "\n",
        "FEATURES = ['dayofyear', 'hour', 'dayofweek', 'quarter', 'month', 'year']\n",
        "TARGET = 'MW'\n",
        "\n",
        "X_train = train[FEATURES]\n",
        "y_train = train[TARGET]\n",
        "\n",
        "X_test = test[FEATURES]\n",
        "y_test = test[TARGET]"
      ],
      "metadata": {
        "id": "223yCyDOoMwv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "reg = xgb.XGBRegressor(base_score=0.5, booster='gbtree',\n",
        "                       n_estimators=1000,\n",
        "                       early_stopping_rounds=50,\n",
        "                       objective='reg:linear',\n",
        "                       max_depth=3,\n",
        "                       learning_rate=0.01)\n",
        "reg.fit(X_train, y_train,\n",
        "        eval_set=[(X_train, y_train), (X_test, y_test)],\n",
        "        verbose=100)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 211
        },
        "id": "xamtoovOozrM",
        "outputId": "cfa1cec9-a1d2-41f4-deb3-c92cde9bf19a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "name 'xgb' is not defined",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-30-97e7e3d0cefa>\u001b[0m in \u001b[0;36m<cell line: 1>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m reg = xgb.XGBRegressor(base_score=0.5, booster='gbtree',\n\u001b[0m\u001b[1;32m      2\u001b[0m                        \u001b[0mn_estimators\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1000\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m                        \u001b[0mearly_stopping_rounds\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m50\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m                        \u001b[0mobjective\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'reg:linear'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m                        \u001b[0mmax_depth\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'xgb' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        " fi = pd.DataFrame(data=reg.feature_importances_,\n",
        "             index=reg.feature_names_in_,\n",
        "             columns=['importance'])\n",
        "fi.sort_values('importance').plot(kind='barh', title='Feature Importance')\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "NhSo_ETNpHck"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.9 Check if US holidays have any correlation or pattern on energy consumption\n",
        "By Yian\n",
        "<br>\n",
        "\n",
        "Hypothesis 1: There is more energy consumption on holiday days"
      ],
      "metadata": {
        "id": "yfcmSUZgsddF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Check Holiday Dates and add is_holiday boolean column\n",
        "df_combined[\"Datetime\"] = pd.to_datetime(df_combined[\"Datetime\"])\n",
        "us_holidays = holidays.US()\n",
        "df_combined[\"is_holiday\"] = df_combined[\"Datetime\"].apply(lambda x: True if x in us_holidays else False)"
      ],
      "metadata": {
        "id": "WpcnJCkPrphh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "correlation = df_combined[\"is_holiday\"].corr(df_combined[\"MW\"])\n",
        "print(f\"Correlation between holidays and energy consumption: {correlation:.2f}\")"
      ],
      "metadata": {
        "id": "O3YOcpUvsvox"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "holiday_consumption = df_combined.groupby(\"is_holiday\")[\"MW\"].mean()\n",
        "print(holiday_consumption)\n"
      ],
      "metadata": {
        "id": "uUNAOrCerpeF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "holiday_consumption.plot(kind=\"bar\", title=\"Average Energy Consumption Holidays vs. Non-holidays\")\n",
        "plt.ylabel(\"Mean Energy Consumption in MW\")\n",
        "plt.xticks([0, 1], [\"Non-holiday\", \"Holiday\"], rotation=0)\n",
        "plt.show()\n"
      ],
      "metadata": {
        "id": "vBdq0yaRspmd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Findings on correlation between average use of energy on holiday days and non-holiday days\n",
        "The correation is -0.01, and from the bar chart we find the average consumption of energy is actually less on holiday days. The low negative correlation shows that there's no significant linear correlation between holidays and energy consumption."
      ],
      "metadata": {
        "id": "taYv_ajOtYJd"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.10 Check specific holiday days (Thanksgiving and Christmas) and one week leading up to these holidays\n",
        "By Yian\n",
        "<br>\n",
        "\n",
        "Check if there's a correlationship between energy consumption one week leading up to the major holidays like Christmas and Thanksgiving, and on the day of these holidays"
      ],
      "metadata": {
        "id": "RFUH0E1RuIoo"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def is_christmas(date):\n",
        "    return date.month == 12 and date.day == 25\n",
        "\n",
        "def is_thanksgiving(date):\n",
        "    return date in us_holidays and us_holidays.get(date) == \"Thanksgiving\"\n",
        "\n",
        "df_combined[\"is_christmas\"] = df_combined[\"Datetime\"].dt.date.apply(is_christmas)\n",
        "df_combined[\"is_thanksgiving\"] = df_combined[\"Datetime\"].dt.date.apply(is_thanksgiving)\n",
        "\n",
        "df_combined[\"is_week_before_christmas\"] = df_combined[\"Datetime\"].dt.date.apply(\n",
        "    lambda x: (pd.Timestamp(year=x.year, month=12, day=18) <= pd.Timestamp(x) < pd.Timestamp(year=x.year, month=12, day=25))\n",
        ")\n",
        "\n",
        "df_combined[\"is_week_before_thanksgiving\"] = df_combined[\"Datetime\"].dt.date.apply(\n",
        "    lambda x: (pd.Timestamp(x) - pd.DateOffset(weeks=1)).date() in us_holidays and us_holidays.get((pd.Timestamp(x) - pd.DateOffset(weeks=1)).date()) == \"Thanksgiving\"\n",
        ")\n"
      ],
      "metadata": {
        "id": "3lpbFWeAttSP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_combined.head()"
      ],
      "metadata": {
        "id": "lTyuekvJzXvh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title calculating correlation\n",
        "\n",
        "christmas_corr = df_combined[\"is_christmas\"].corr(df_combined[\"MW\"])\n",
        "thanksgiving_corr = df_combined[\"is_thanksgiving\"].corr(df_combined[\"MW\"])\n",
        "week_before_christmas_corr = df_combined[\"is_week_before_christmas\"].corr(df_combined[\"MW\"])\n",
        "week_before_thanksgiving_corr = df_combined[\"is_week_before_thanksgiving\"].corr(df_combined[\"MW\"])\n",
        "\n",
        "print(f\"Correlation on Christmas: {christmas_corr:.4f}\")\n",
        "print(f\"Correlation on Thanksgiving: {thanksgiving_corr:.4f}\")\n",
        "print(f\"Correlation 1 week before Christmas: {week_before_christmas_corr:.4f}\")\n",
        "print(f\"Correlation 1 week before Thanksgiving: {week_before_thanksgiving_corr:.4f}\")"
      ],
      "metadata": {
        "id": "H56HmLExufKl"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Visualization on average daily consumption of energy one week before Christmas and On Christmas\n",
        "# aggregate data by daily mean instead of hourly\n",
        "df_combined['Date'] = df_combined['Datetime'].dt.date\n",
        "daily_consumption = df_combined.groupby('Date')['MW'].mean().reset_index()\n",
        "\n",
        "# check energy consumption pattern a week before christmas and on christmas\n",
        "daily_consumption['is_christmas'] = daily_consumption['Date'].apply(lambda x: is_christmas(pd.Timestamp(x)))\n",
        "daily_consumption['is_week_before_christmas'] = daily_consumption['Date'].apply(\n",
        "    lambda x: pd.Timestamp(x) >= pd.Timestamp(x.year, 12, 18) and pd.Timestamp(x) < pd.Timestamp(x.year, 12, 25)\n",
        ")\n",
        "\n",
        "plt.figure(figsize=(14, 6))\n",
        "plt.scatter(daily_consumption[daily_consumption['is_christmas']]['Date'],\n",
        "            daily_consumption[daily_consumption['is_christmas']]['MW'],\n",
        "            color='red', label=\"Christmas\", s=50)\n",
        "plt.scatter(daily_consumption[daily_consumption['is_week_before_christmas']]['Date'],\n",
        "            daily_consumption[daily_consumption['is_week_before_christmas']]['MW'],\n",
        "            color='orange', label=\"Week Before Christmas\", s=50)\n",
        "plt.plot(daily_consumption['Date'], daily_consumption['MW'], label=\"Daily Avg Energy Consumption\", color='blue',alpha=0.5)\n",
        "\n",
        "\n",
        "plt.title(\"Daily Average Energy Consumption Before and On Christmas\")\n",
        "plt.xlabel(\"Date\")\n",
        "plt.ylabel(\"Average MW\")\n",
        "plt.legend()\n",
        "plt.grid()\n",
        "plt.show()\n"
      ],
      "metadata": {
        "id": "cuihneHvvR3W"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Visualization on daily average energy consumption one week before Thanksgiving and on Thanksgiving day\n",
        "daily_consumption['is_thanksgiving'] = daily_consumption['Date'].apply(lambda x: is_thanksgiving(pd.Timestamp(x)))\n",
        "daily_consumption['is_week_before_thanksgiving'] = daily_consumption['Date'].apply(\n",
        "    lambda x: (pd.Timestamp(x) - pd.DateOffset(weeks=1)).date() in us_holidays and\n",
        "              us_holidays.get((pd.Timestamp(x) - pd.DateOffset(weeks=1)).date()) == \"Thanksgiving\"\n",
        ")\n",
        "\n",
        "plt.figure(figsize=(14, 6))\n",
        "\n",
        "plt.scatter(daily_consumption[daily_consumption['is_thanksgiving']]['Date'],\n",
        "            daily_consumption[daily_consumption['is_thanksgiving']]['MW'],\n",
        "            color='red', label=\"Thanksgiving\", s=50)\n",
        "plt.scatter(daily_consumption[daily_consumption['is_week_before_thanksgiving']]['Date'],\n",
        "            daily_consumption[daily_consumption['is_week_before_thanksgiving']]['MW'],\n",
        "            color='orange', label=\"Week Before Thanksgiving\", s=100)\n",
        "\n",
        "plt.plot(daily_consumption['Date'], daily_consumption['MW'], label=\"Daily Avg Energy Consumption\", color='blue',alpha=0.5)\n",
        "\n",
        "plt.title(\"Daily Average Energy Consumption Before and On Thanksgiving\")\n",
        "plt.xlabel(\"Date\")\n",
        "plt.ylabel(\"Average MW\")\n",
        "plt.legend()\n",
        "plt.grid()\n",
        "plt.show()\n"
      ],
      "metadata": {
        "id": "IzlIwFspv4Pf"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "christmas_data = daily_consumption[['MW', 'is_christmas', 'is_week_before_christmas']]\n",
        "christmas_corr = christmas_data.corr(method='pearson')\n",
        "thanksgiving_data = daily_consumption[['MW', 'is_thanksgiving', 'is_week_before_thanksgiving']]\n",
        "thanksgiving_corr = thanksgiving_data.corr(method='pearson')\n",
        "\n",
        "fig, axes = plt.subplots(1, 2, figsize=(16, 6))\n",
        "\n",
        "sns.heatmap(christmas_corr, annot=True, cmap='coolwarm', linewidths=0.5, ax=axes[0])\n",
        "axes[0].set_title(\"Pearson Correlation (Christmas)\")\n",
        "\n",
        "sns.heatmap(thanksgiving_corr, annot=True, cmap='coolwarm', linewidths=0.5, ax=axes[1])\n",
        "axes[1].set_title(\"Pearson Correlation (Thanksgiving)\")\n",
        "\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "UdOk65R7LWzM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Findings on correlation between major winter holidays\n",
        "The correlation between holidays and a week leading up to holidays and average energy daily consumption seem very weak, suggesting there is no linear correlation. Hypothesis 1 is False and the relevance is low."
      ],
      "metadata": {
        "id": "htcPthceymH2"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.11 Check correlation between seasons and energy consumption\n",
        "By Yian\n",
        "<br>\n",
        "Hypothesis 2: Different seasons could have different energy consumption needs. Seasons with higher or lower temperatures require more energy use."
      ],
      "metadata": {
        "id": "yYxELty82H_h"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Categorize the seasons by month\n",
        "def get_season(date):\n",
        "    month = date.month\n",
        "    if month in [12, 1, 2]:\n",
        "        return 'Winter'\n",
        "    elif month in [3, 4, 5]:\n",
        "        return 'Spring'\n",
        "    elif month in [6, 7, 8]:\n",
        "        return 'Summer'\n",
        "    elif month in [9, 10, 11]:\n",
        "        return 'Autumn'\n",
        "    return None\n",
        "\n",
        "daily_consumption['season'] = daily_consumption['Date'].apply(lambda x: get_season(pd.Timestamp(x)))\n",
        "\n"
      ],
      "metadata": {
        "id": "IoEnS2Rl1Hoo"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "daily_consumption.head()"
      ],
      "metadata": {
        "id": "03_MK3oC1myb"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "seasonal_avg = daily_consumption.groupby('season')['MW'].mean().sort_index()\n",
        "\n",
        "print(seasonal_avg)\n"
      ],
      "metadata": {
        "id": "lAc3P5F32cZQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "plt.figure(figsize=(10, 6))\n",
        "seasonal_avg.plot(kind='bar', color=['skyblue', 'lightgreen', 'salmon', 'orange'])\n",
        "plt.title(\"Average Energy Consumption by Season\")\n",
        "plt.xlabel(\"Season\")\n",
        "plt.ylabel(\"Average MW\")\n",
        "plt.xticks(rotation=0)\n",
        "plt.grid(axis='y')\n",
        "plt.show()\n"
      ],
      "metadata": {
        "id": "kd_dgx5W2gsk"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "daily_consumption.head()"
      ],
      "metadata": {
        "id": "Zxe_ZvOx5p02"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df = df_combined.copy()"
      ],
      "metadata": {
        "id": "9m__Yw0B7XSn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df['Datetime'] = pd.to_datetime( df['Datetime'] )\n",
        "df['Year'] = df['Datetime'].dt.year"
      ],
      "metadata": {
        "id": "M1ArA8Vw6jEj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df['season'] = df['Date'].apply(lambda x: get_season(pd.Timestamp(x)))\n"
      ],
      "metadata": {
        "id": "5VG9HP8o7mHk"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.head()"
      ],
      "metadata": {
        "id": "LDlRuZDl72i7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "aux2 = df[['Year', 'season', 'MW']].groupby(['Year', 'season']).sum().reset_index()\n",
        "\n",
        "plt.figure(figsize=(14, 8))\n",
        "# sns.color_palette(\"hls\", 8)\n",
        "\n",
        "sns.barplot(x='Year', y='MW', palette='husl',hue='season', data=aux2)\n",
        "plt.title(\"Energy Consumption by Year and Season\")\n",
        "plt.xlabel(\"Year\")\n",
        "plt.ylabel(\"Total MW Energy Consumption\")\n",
        "plt.xticks(rotation=45)\n",
        "plt.legend(title=\"Season\")\n",
        "plt.grid(axis='y', linestyle='--', alpha=1)\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "Q1ne5HBz76HY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Findings on different seasons having different energy consumption needs\n",
        "The yearly aggregate bar chart and average bar chart show that in summer seasons the energy consumption in the US is usually highest. The second highest is normally winter or fall seasons. The pattern is consistent through the years, therefore hypothesis 2 is true.\n"
      ],
      "metadata": {
        "id": "c6GS1dj_-hkW"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 1.12 Visualize how hour of the day affects energy expenditure\n",
        "By Yian\n"
      ],
      "metadata": {
        "id": "0jwKpsibNug7"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df['hour_of_day'] = df['Datetime'].dt.hour\n",
        "aux3 = df[['hour_of_day', 'MW']].groupby( 'hour_of_day' ).sum().reset_index()\n",
        "fig, axes = plt.subplots(1, 2, figsize=(16, 6))\n",
        "sns.barplot( x='hour_of_day', y='MW', data=aux3, ax=axes[0])\n",
        "sns.heatmap( aux3.corr( method='pearson' ), annot=True, ax=axes[1] );"
      ],
      "metadata": {
        "id": "jeUE7tzXJRAw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df['day_of_month'] = df['Datetime'].dt.day\n",
        "df['day_of_week'] = df['Datetime'].dt.dayofweek\n",
        "df['Month'] = df['Datetime'].dt.month"
      ],
      "metadata": {
        "id": "dveePH71T-GY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 2. Feature Selection and Data Preparation\n",
        "By Yian\n"
      ],
      "metadata": {
        "id": "E7q4fxK4P2E4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.ensemble import RandomForestRegressor\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.metrics import r2_score, mean_squared_error\n",
        "from sklearn.base import BaseEstimator, TransformerMixin\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n"
      ],
      "metadata": {
        "id": "fuuKp3cFDku8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.columns"
      ],
      "metadata": {
        "id": "y66ivZpBdhwc"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df2 = df.copy()"
      ],
      "metadata": {
        "id": "_r3Ke-Ocp7bO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df2.head()"
      ],
      "metadata": {
        "id": "kn384b-Dei5d"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2.1 Synchronize the time period for the companies since the datasets start and end in different years/ date ranges\n",
        "\n",
        "By Yian"
      ],
      "metadata": {
        "id": "fKMVTr7tMBsv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the min and max dates of each datasets of different companies to see if there are overlaps\n",
        "min_max = []\n",
        "for i in df2['company'].unique():\n",
        "    min_max.append( [ i, df2[df2['company']==i]['Date'].min(), df2[df2['company']==i]['Date'].max() ] )\n",
        "\n",
        "mm = pd.DataFrame( columns=[ 'electric_company', 'min_date', 'max_date' ], data=min_max )\n",
        "mm"
      ],
      "metadata": {
        "id": "lPy88Rsoflpb"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        " Since there are no same min and max dates for all the electrical companies, for our training data we will use all data points from earliest data (min_date) up to 1 year before last date (max_date)\n",
        " to ensure the training data captures the long history but excludes the most recent year.\n",
        " For our test set we will include the most recent data points from cut date (1 year before max_date) up to the latest date (max_date).\n"
      ],
      "metadata": {
        "id": "OvjNFRiTlM5D"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import datetime\n",
        "cut_dates = []\n",
        "\n",
        "# get the cut dates for all companies, since all date ranges of the datasets of companies are different\n",
        "for company in mm['electric_company'].unique():\n",
        "    max_date = pd.to_datetime(mm[mm['electric_company'] == company]['max_date'].values[0])\n",
        "\n",
        "    cut_date = max_date - datetime.timedelta(days=12 * 30)  # approximate 12 months as 360 days\n",
        "\n",
        "    cut_dates.append([company, cut_date])\n",
        "\n",
        "cutoff_df = pd.DataFrame(cut_dates, columns=['electric_company', 'cut_date'])\n",
        "cutoff_df"
      ],
      "metadata": {
        "id": "YY1PQnIroTSX"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def add_company_id(X):\n",
        "  \"\"\"\n",
        "  encodes categorical company names to numerical values\n",
        "  \"\"\"\n",
        "  X['company_id'] = X['company'].astype('category').cat.codes\n",
        "\n",
        "  return X"
      ],
      "metadata": {
        "id": "WKrVlg8LYM6p"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 2.1.1 Function that synchronizes our train and test data\n",
        "The training data will contain data points from the earliest data up to 1 year before last date. The test set will include the most recent data points from cut date (1 year before before the latest date)."
      ],
      "metadata": {
        "id": "jcrJinvvYhqh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def synchronize_data(df2, cutoff_df):\n",
        "  \"\"\"\n",
        "  synchronizes all energy consumption data of different companies to established cut-off point for train-test split\n",
        "  \"\"\"\n",
        "  df2 = add_company_id(df2.copy())\n",
        "\n",
        "  train_data = []\n",
        "  test_data = []\n",
        "\n",
        "  for company in cutoff_df['electric_company']:\n",
        "      # cut date for each company\n",
        "      cut_date = cutoff_df[cutoff_df['electric_company'] == company]['cut_date'].values[0]\n",
        "\n",
        "      company_data = df2[df2['company'] == company]\n",
        "\n",
        "      train = company_data[company_data['Datetime'] < cut_date]\n",
        "      test = company_data[company_data['Datetime'] >= cut_date]\n",
        "\n",
        "      train_data.append(train)\n",
        "      test_data.append(test)\n",
        "\n",
        "  train_df = pd.concat(train_data)\n",
        "  test_df = pd.concat(test_data)\n",
        "  return train_df, test_df"
      ],
      "metadata": {
        "id": "pea_GRSuYB0Q"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2.2 Creating Data Pipeline and Applying Feature Transformation\n",
        "By Yian\n",
        "<br>\n",
        "<br>\n",
        "### 2.2.1 Encoding Cyclical Features\n",
        "\n",
        "<br>\n",
        "A common method to encode cyclical data is to transform data into two dimensions using sine and cosine transformation.\n",
        "We want to transform the original features into cyclical features (hour of day, day of week, day of month, season) to capture the cyclical nature of these variables. Since we're working with time series data, features like hour of the day, day of the week, and seasons etc. have a cyclical pattern. By applying sine and cosine transformations we project these values onto a circle for our models to learn the cyclical natures. Sine captures the y-axis information and cosine captures the x-axis information.\n",
        "<br>\n",
        "We perform the following transformations:\n",
        "\\begin{align*}\n",
        "x_{\\sin} &= \\sin\\left(\\frac{2 \\cdot \\pi \\cdot x}{\\max(x)}\\right) \\\\\n",
        "x_{\\cos} &= \\cos\\left(\\frac{2 \\cdot \\pi \\cdot x}{\\max(x)}\\right)\n",
        "\\end{align*}\n",
        "<br>\n",
        "<br>\n",
        "\n",
        "### 2.2.2 Building Preprocessing Pipeline\n",
        "For our data pipeline, we apply feature transformations such as encoding features to cyclical patterns using sine and cosine. We also encode the seasons and holiday features.\n",
        "<br>\n",
        "No further normalization is needed for this data since transforming the data to cyclical features (with sine and cosine) already normalizes them between -1 and 1.\n"
      ],
      "metadata": {
        "id": "1hfXbHBzMZEP"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class CustomPreprocessor(BaseEstimator, TransformerMixin):\n",
        "    def __init__(self):\n",
        "      pass\n",
        "\n",
        "    def fit(self, X, y=None):\n",
        "        return self\n",
        "\n",
        "    def encode_seasons_holidays (self, X):\n",
        "      \"\"\"\n",
        "      encodes seasons and holiday values\n",
        "      \"\"\"\n",
        "      # Mapping the seasons categorical variable to ordinal values because there is ordinal relationship\n",
        "      ordinal_encode_season = {'Winter': 1, 'Spring': 2, 'Summer': 3, 'Autumn': 4}\n",
        "      X['season'] = X['season'].map(ordinal_encode_season)\n",
        "\n",
        "      # One-hot encode holidays (change to binary values)\n",
        "      X['is_holiday'] = X['is_holiday'].astype(int)\n",
        "\n",
        "      return X\n",
        "\n",
        "    def transform(self, X):\n",
        "      \"\"\"\n",
        "      returns X with encoded features\n",
        "      \"\"\"\n",
        "\n",
        "      X = self.encode_seasons_holidays(X.copy())\n",
        "\n",
        "      # Sine-cosine transformation for cyclical features\n",
        "      # hour of day range from 0 to 23, so this scales our hour of the day to fit in circle 0 to 2pi radians\n",
        "      X['hour_of_day_sin'] = X['hour_of_day'].apply(lambda x: np.sin(x * (2. * np.pi / 24.)))\n",
        "      X['hour_of_day_cos'] = X['hour_of_day'].apply(lambda x: np.cos(x * (2. * np.pi / 24.)))\n",
        "\n",
        "      # split seasons to 4 parts\n",
        "      X['season_sin'] = X['season'].apply(lambda x: np.sin(x * (2. * np.pi / 4.)))\n",
        "      X['season_cos'] = X['season'].apply(lambda x: np.cos(x * (2. * np.pi / 4.)))\n",
        "\n",
        "      X['day_of_week_sin'] = np.sin(X['day_of_week'] * (2. * np.pi / 7.))\n",
        "      X['day_of_week_cos'] = np.cos(X['day_of_week'] * (2. * np.pi / 7.))\n",
        "\n",
        "      X['day_of_month_sin'] = np.sin(X['day_of_month'] * (2. * np.pi / 31.))\n",
        "      X['day_of_month_cos'] = np.cos(X['day_of_month'] * (2. * np.pi / 31.))\n",
        "\n",
        "      # Drop unwanted columns\n",
        "      cols_drop = ['hour_of_day','is_thanksgiving','is_christmas','is_week_before_christmas','is_week_before_thanksgiving', 'season', 'day_of_week', 'day_of_month']\n",
        "      X = X.drop(columns=cols_drop, errors='ignore')\n",
        "\n",
        "      return X\n"
      ],
      "metadata": {
        "id": "g5S5fyugEx9r"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 3. Split dataframe to train and test dataset\n",
        "By Yian\n",
        "\n"
      ],
      "metadata": {
        "id": "Kpp1WluceuuR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "features = [\n",
        "    'hour_of_day_sin', 'hour_of_day_cos',\n",
        "    'season_sin', 'season_cos',\n",
        "    'day_of_week_sin', 'day_of_week_cos',\n",
        "    'day_of_month_sin', 'day_of_month_cos',\n",
        "    'company_id'\n",
        "]\n",
        "\n",
        "target = 'MW'\n",
        "\n",
        "preprocessor = CustomPreprocessor()\n",
        "df_transformed = preprocessor.transform(df2)\n",
        "train_df, test_df = synchronize_data(df_transformed, cutoff_df)\n",
        "\n",
        "# get select features\n",
        "X_train, y_train = train_df[features], train_df[target]\n",
        "X_test, y_test = test_df[features], test_df[target]\n",
        "\n",
        "print(f\"X_train shape: {X_train.shape}\")\n",
        "print(f\"y_train shape: {y_train.shape}\")\n",
        "print(f\"X_test shape: {X_test.shape}\")\n",
        "print(f\"y_test shape: {y_test.shape}\")\n"
      ],
      "metadata": {
        "id": "zhtoLu9XMMbb"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train.head()"
      ],
      "metadata": {
        "id": "rj8rOP-ZGSu5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 4. Random Forest Model\n",
        "By Yian\n"
      ],
      "metadata": {
        "id": "NhYLNz9K_k0w"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.ensemble import RandomForestRegressor\n",
        "from sklearn.metrics import r2_score, mean_squared_error\n",
        "\n",
        "\n",
        "initial_rf_model = RandomForestRegressor(n_estimators=100, random_state=42)\n",
        "initial_rf_model.fit(X_train, y_train)\n"
      ],
      "metadata": {
        "id": "DCE3-8-2Igeh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#@title Predict Training Data\n",
        "y_train_pred = initial_rf_model.predict(X_train)\n",
        "\n",
        "y_test_pred = initial_rf_model.predict(X_test)\n"
      ],
      "metadata": {
        "id": "DAeQdhhzs0aD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 4.1 Model Evaluation with R-Squared and RMSE\n",
        "By Yian\n",
        "<br>\n",
        "We evaluate with R-squared $R^2 = 1 - \\frac{\\sum_{i=1}^{n} (y_i - \\hat{y}_i)^2}{\\sum_{i=1}^{n} (y_i - \\bar{y})^2}$, where y_i is the target value and y_hat is the predicted value, to check how well the model fits the data (1.0 is perfect fit). We also use Root Mean Sqaured Error (RMSE) $\\text{RMSE} = \\sqrt{\\frac{1}{n} \\sum_{i=1}^{n} (y_i - \\hat{y}_i)^2}\n",
        "$\n",
        "to measure the standard deviation of residuals (predition errors). The lower the RMSE means less error, which indicates better fit."
      ],
      "metadata": {
        "id": "q5DwP7I0s507"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_r2 = r2_score(y_train, y_train_pred)\n",
        "test_r2 = r2_score(y_test, y_test_pred)\n",
        "\n",
        "train_rmse = np.sqrt(mean_squared_error(y_train, y_train_pred))\n",
        "test_rmse = np.sqrt(mean_squared_error(y_test, y_test_pred))\n",
        "\n",
        "print(f\"Training R² Score: {train_r2:.4f}\")\n",
        "print(f\"Test R² Score: {test_r2:.4f}\")\n",
        "print(f\"Training RMSE: {train_rmse:.2f}\")\n",
        "print(f\"Test RMSE: {test_rmse:.2f}\")\n"
      ],
      "metadata": {
        "id": "gsdux9Q8tZID"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 4.2 Check Feature Importance of the Random Forest Model\n",
        "Yian"
      ],
      "metadata": {
        "id": "OVmtqGy2t0xs"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "feature_importance = initial_rf_model.feature_importances_\n",
        "\n",
        "importance_df = pd.DataFrame({'Feature': features, 'Importance': feature_importance})\n",
        "importance_df = importance_df.sort_values(by='Importance', ascending=False)\n",
        "\n",
        "plt.figure(figsize=(10, 6))\n",
        "plt.barh(importance_df['Feature'], importance_df['Importance'])\n",
        "plt.xlabel(\"Importance\")\n",
        "plt.title(\"Feature Importance in Random Forest Regressor\")\n",
        "plt.show()\n"
      ],
      "metadata": {
        "id": "Aq2XvG8xt3qD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "The high feature importance of 'company_id' in the Random Forest model is likely because the company_id acts as a unique categorical identifier of the source of the data. Since different electric companies in the dataset have unique patterns of energy consumption due to geographical, environmental, and operational differences. These companies are in different regions and have distinct seasonal behaviors and energy demands. As a result, the 'company_id' could help the model easily differentiate and make better predictions based on the company.\n"
      ],
      "metadata": {
        "id": "6pi7l2ST9qiz"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 4.3 Create Feature Selector Transformer\n",
        "By Yian\n",
        "<br>\n",
        "Build top k feature selector based on Random Forest base model feature importance to add to our data pipeline"
      ],
      "metadata": {
        "id": "FjS2ITlWHAFG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def indices_of_top_k(arr, k):\n",
        "    return np.sort(np.argpartition(np.array(arr), -k)[-k:])\n",
        "\n",
        "class TopFeatureSelector(BaseEstimator, TransformerMixin):\n",
        "    def __init__(self, feature_importances, k):\n",
        "        self.feature_importances = feature_importances\n",
        "        self.k = k\n",
        "\n",
        "    def fit(self, X, y=None):\n",
        "        self.feature_indices_ = indices_of_top_k(self.feature_importances, self.k)\n",
        "        return self\n",
        "\n",
        "    def transform(self, X):\n",
        "        return X.iloc[:, self.feature_indices_]\n"
      ],
      "metadata": {
        "id": "pbdYxRA3G-0O"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 4.4 Putting together the data pipeline for feature selector and running RF model after feature selection on initial model\n",
        "By Yian"
      ],
      "metadata": {
        "id": "OhTjYH0_TEfh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# pipeline after adding feature selector\n",
        "\n",
        "top_k_features = 5\n",
        "\n",
        "data_pipeline = Pipeline([\n",
        "    ('feature_selector', TopFeatureSelector(feature_importance, top_k_features)),\n",
        "    ('regressor', RandomForestRegressor(n_estimators=100, random_state=42))\n",
        "])\n",
        "\n",
        "data_pipeline.fit(X_train, y_train)\n",
        "\n",
        "# predictions\n",
        "y_train_pred = data_pipeline.predict(X_train)\n",
        "y_test_pred = data_pipeline.predict(X_test)"
      ],
      "metadata": {
        "id": "qX8yF_nmHVjO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 4.5 Evaluation on RF Model after top 5 feature selection\n",
        "By Yian"
      ],
      "metadata": {
        "id": "MIYxPD-HRXl4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_r2 = r2_score(y_train, y_train_pred)\n",
        "test_r2 = r2_score(y_test, y_test_pred)\n",
        "\n",
        "train_rmse = np.sqrt(mean_squared_error(y_train, y_train_pred))\n",
        "test_rmse = np.sqrt(mean_squared_error(y_test, y_test_pred))\n",
        "\n",
        "print(f\"Training R² Score: {train_r2:.4f}\")\n",
        "print(f\"Test R² Score: {test_r2:.4f}\")\n",
        "print(f\"Training RMSE: {train_rmse:.2f}\")\n",
        "print(f\"Test RMSE: {test_rmse:.2f}\")"
      ],
      "metadata": {
        "id": "ezpypCOWRgA7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Conclusion on Random Forest Model\n",
        "By Yian\n",
        "\n",
        "The Random Forest initial model achieves a fairly high score with an RMSE of 1987.45 on the test set and 0.962 for the R^2. The Random Forest model after feature selections indicates even more robust performance of 1752.68 on the test set (lower prediction error) and R^2 score of 0.97, which demonstrates that the selecting only the top k important features helps reduce the feature space (dimensionality of data) and model complexity, which in turn helps the model generalize even better."
      ],
      "metadata": {
        "id": "GVGRaf-paMK_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# 5 .RNN Model\n",
        "\n",
        "By Hogan"
      ],
      "metadata": {
        "id": "bWD2f7qIyF0y"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "import os\n",
        "import sklearn.preprocessing\n",
        "from sklearn.metrics import r2_score\n",
        "from keras.layers import Dense,Dropout,SimpleRNN,LSTM\n",
        "from keras.models import Sequential"
      ],
      "metadata": {
        "id": "OBHx35IC-l9b"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Plotting hourly energy usage:\n",
        "AEP = pd.read_csv(os.path.join(path, \"AEP_hourly.csv\"), index_col=[0], parse_dates=[0])\n",
        "\n",
        "mau = [\"#F8766D\", \"#D39200\", \"#93AA00\", \"#00BA38\", \"#00C19F\", \"#00B9E3\", \"#619CFF\", \"#DB72FB\"]\n",
        "bieudo = AEP.plot(style='.',figsize=(15,5), color=mau[0], title='AEP')\n",
        "\n",
        "#Data transformation\n",
        "def create_features(df, label=None):\n",
        "    df = df.copy()\n",
        "    df['date'] = df.index\n",
        "    df['hour'] = df['date'].dt.hour\n",
        "    df['dayofweek'] = df['date'].dt.dayofweek\n",
        "    df['quarter'] = df['date'].dt.quarter\n",
        "    df['month'] = df['date'].dt.month\n",
        "    df['year'] = df['date'].dt.year\n",
        "    df['dayofyear'] = df['date'].dt.dayofyear\n",
        "    df['dayofmonth'] = df['date'].dt.day\n",
        "\n",
        "    X = df[['hour', 'dayofweek', 'quarter', 'month', 'year',\n",
        "            'dayofyear', 'dayofmonth']]\n",
        "    if label:\n",
        "        y = df[label]\n",
        "        return X, y\n",
        "    return X\n",
        "\n",
        "X, y = create_features(AEP, label='AEP_MW')\n",
        "features_and_target = pd.concat([X, y], axis=1)\n",
        "print(features_and_target)\n",
        "plt.show()\n",
        "\n",
        "plt.figure(figsize=(15,6))\n",
        "data_csv = AEP.dropna()\n",
        "dataset = data_csv.values\n",
        "dataset = dataset.astype('float32')\n",
        "max_value = np.max(dataset)\n",
        "min_value = np.min(dataset)\n",
        "scalar = max_value - min_value\n",
        "dataset = list(map(lambda x: (x-min_value) / scalar, dataset))\n",
        "plt.plot(dataset)\n",
        "print(max_value, min_value)"
      ],
      "metadata": {
        "id": "gLLCj-a7-pL7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#choosing DOM_hourly.csv data for analysis\n",
        "fpath='./dataset/DOM_hourly.csv'\n",
        "\n",
        "#Let's use datetime(2012-10-01 12:00:00,...) as index instead of numbers(0,1,...)\n",
        "#This will be helpful for further data analysis as we are dealing with time series data\n",
        "df = pd.read_csv(os.path.join(path, \"DOM_hourly.csv\"), index_col='Datetime', parse_dates=['Datetime'])\n",
        "df.head()\n",
        "\n",
        "#checking missing data\n",
        "df.isna().sum()\n",
        "\n",
        "#Data visualization\n",
        "\n",
        "df.plot(figsize=(16,4),legend=True)\n",
        "\n",
        "plt.title('DOM hourly power consumption data - BEFORE NORMALIZATION')\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "CxAxkDJ6-rB5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#Normalize DOM hourly power consumption data\n",
        "\n",
        "def normalize_data(df):\n",
        "    scaler = sklearn.preprocessing.MinMaxScaler()\n",
        "    df['DOM_MW']=scaler.fit_transform(df['DOM_MW'].values.reshape(-1,1))\n",
        "    return df\n",
        "\n",
        "df_norm = normalize_data(df)\n",
        "df_norm.shape\n",
        "\n",
        "#Visualize data after normalization\n",
        "\n",
        "df_norm.plot(figsize=(16,4),legend=True)\n",
        "\n",
        "plt.title('DOM hourly power consumption data - AFTER NORMALIZATION')\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "zYF48lQP-thy"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# train data for deep learning models\n",
        "\n",
        "def load_data(stock, seq_len):\n",
        "    X_train = []\n",
        "    y_train = []\n",
        "    for i in range(seq_len, len(stock)):\n",
        "        X_train.append(stock.iloc[i - seq_len: i, 0])\n",
        "        y_train.append(stock.iloc[i, 0])\n",
        "\n",
        "    # 1 last 6189 days are going to be used in test\n",
        "    X_test = X_train[110000:]\n",
        "    y_test = y_train[110000:]\n",
        "\n",
        "    # 2 first 110000 days are going to be used in training\n",
        "    X_train = X_train[:110000]\n",
        "    y_train = y_train[:110000]\n",
        "\n",
        "    # 3 convert to numpy array\n",
        "    X_train = np.array(X_train)\n",
        "    y_train = np.array(y_train)\n",
        "\n",
        "    X_test = np.array(X_test)\n",
        "    y_test = np.array(y_test)\n",
        "\n",
        "    # 4 reshape data to input into RNN models\n",
        "    X_train = np.reshape(X_train, (110000, seq_len, 1))\n",
        "\n",
        "    X_test = np.reshape(X_test, (X_test.shape[0], seq_len, 1))\n",
        "\n",
        "    return [X_train, y_train, X_test, y_test]"
      ],
      "metadata": {
        "id": "_1HNYay8-vmE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#create train, test data\n",
        "seq_len = 20 #choose sequence length\n",
        "\n",
        "X_train, y_train, X_test, y_test = load_data(df, seq_len)\n",
        "\n",
        "print('X_train.shape = ',X_train.shape)\n",
        "print('y_train.shape = ', y_train.shape)\n",
        "print('X_test.shape = ', X_test.shape)\n",
        "print('y_test.shape = ',y_test.shape)"
      ],
      "metadata": {
        "id": "ZKbKWFZy-xKD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "By Pranay\n",
        "We want to compare different models using the performance metrics , here we are using a Simple RNN as a base to start with, The performance metrics we used is R2 score and by adjusting the number of layers and the dropout rate we want to increase the performance."
      ],
      "metadata": {
        "id": "wWWyg4fia6cG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#RNN model\n",
        "rnn_model = Sequential()\n",
        "rnn_model.add(SimpleRNN(40,activation=\"tanh\",return_sequences=True, input_shape=(X_train.shape[1],1)))\n",
        "rnn_model.add(Dropout(0.15))\n",
        "rnn_model.add(SimpleRNN(40,activation=\"tanh\",return_sequences=True))\n",
        "rnn_model.add(Dropout(0.2))\n",
        "rnn_model.add(SimpleRNN(40,activation=\"tanh\",return_sequences=True))\n",
        "rnn_model.add(Dropout(0.2))\n",
        "rnn_model.add(SimpleRNN(40,activation=\"tanh\",return_sequences=False))\n",
        "rnn_model.add(Dropout(0.25))\n",
        "rnn_model.add(Dense(1))\n",
        "rnn_model.summary()\n",
        "rnn_model.compile(optimizer=\"adam\",loss=\"MSE\")\n",
        "rnn_model.fit(X_train, y_train, epochs=10, batch_size=1000)"
      ],
      "metadata": {
        "id": "_FaX0so_-y3c"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#r2 score for the values predicted by the above trained SIMPLE RNN model\n",
        "\n",
        "rnn_predictions = rnn_model.predict(X_test)\n",
        "rnn_score = r2_score(y_test,rnn_predictions)\n",
        "print(\"R2 Score of RNN model = \",rnn_score)"
      ],
      "metadata": {
        "id": "bYpEtGxA-0Wx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# compare the actual values vs predicted values by plotting a graph\n",
        "\n",
        "def plot_predictions(test, predicted, title):\n",
        "    plt.figure(figsize=(16, 4))\n",
        "    plt.plot(test, color='blue', label='Actual power consumption data')\n",
        "    plt.plot(predicted, alpha=0.7, color='orange', label='Predicted power consumption data')\n",
        "    plt.title(title)\n",
        "    plt.xlabel('Time')\n",
        "    plt.ylabel('Normalized power consumption scale')\n",
        "    plt.legend()\n",
        "    plt.show()\n",
        "plot_predictions(y_test, rnn_predictions, \"Predictions made by simple RNN model\")"
      ],
      "metadata": {
        "id": "zkbIj0x9-2EJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#train model for LSTM\n",
        "\n",
        "lstm_model = Sequential()\n",
        "lstm_model.add(LSTM(40,activation=\"tanh\",return_sequences=True, input_shape=(X_train.shape[1],1)))\n",
        "lstm_model.add(Dropout(0.15))\n",
        "lstm_model.add(LSTM(40,activation=\"tanh\",return_sequences=True))\n",
        "lstm_model.add(Dropout(0.15))\n",
        "lstm_model.add(LSTM(40,activation=\"tanh\",return_sequences=False))\n",
        "lstm_model.add(Dropout(0.15))\n",
        "lstm_model.add(Dense(1))\n",
        "lstm_model.summary()\n",
        "lstm_model.compile(optimizer=\"adam\",loss=\"MSE\")\n",
        "lstm_model.fit(X_train, y_train, epochs=10, batch_size=1000)"
      ],
      "metadata": {
        "id": "2lRlD-_m-4uu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#r2 score for the values predicted by the above trained LSTM model\n",
        "lstm_predictions = lstm_model.predict(X_test)\n",
        "lstm_score = r2_score(y_test, lstm_predictions)\n",
        "print(\"R^2 Score of LSTM model = \",lstm_score)"
      ],
      "metadata": {
        "id": "oe6K8Dav-6vp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#actual values vs predicted values by plotting a graph\n",
        "plot_predictions(y_test, lstm_predictions, \"Predictions made by LSTM model\")"
      ],
      "metadata": {
        "id": "tbgL_NFG-8lE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#RNN, LSTM model by plotting data in a single graph\n",
        "plt.figure(figsize=(15,8))\n",
        "plt.plot(y_test, c=\"orange\", linewidth=3, label=\"Original values\")\n",
        "plt.plot(lstm_predictions, c=\"red\", linewidth=3, label=\"LSTM predictions\")\n",
        "plt.plot(rnn_predictions, alpha=0.5, c=\"blue\", linewidth=3, label=\"RNN predictions\")\n",
        "plt.legend()\n",
        "plt.title(\"Predictions vs actual data\", fontsize=20)\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "GgVfy3ya--qa"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}