Anomalies_Detection.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.10"
    },
    "colab": {
      "name": "Anomalies_Detection.ipynb",
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/FreelancerFel/Anomalies_Detection/blob/master/Anomalies_Detection.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-gCPY0Wsij36"
      },
      "source": [
        "# Understanding Anomalies in Time-Series - Is something going on?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SkwcEmETjTtr",
        "outputId": "dc6c39a4-8c82-4a47-b948-3702432c0ee5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "!pip install burst_detection\n",
        "!pip install torch\n",
        "!pip install torchvision\n",
        "!apt-get install subversion\n",
        "!svn export https://github.com/FreelancerFel/Anomalies_Detection/trunk/datasets\n",
        "\n",
        "# Source for burst_detection implementation\n",
        "# https://github.com/nmarinsek/burst_detection\n",
        "# Source for Numenta Anomaly Benchmark\n",
        "# https://github.com/numenta/NAB"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: burst_detection in /usr/local/lib/python3.6/dist-packages (0.1.3)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from burst_detection) (1.18.5)\n",
            "Requirement already satisfied: pandas in /usr/local/lib/python3.6/dist-packages (from burst_detection) (1.1.4)\n",
            "Requirement already satisfied: mpmath in /usr/local/lib/python3.6/dist-packages (from burst_detection) (1.1.0)\n",
            "Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.6/dist-packages (from pandas->burst_detection) (2018.9)\n",
            "Requirement already satisfied: python-dateutil>=2.7.3 in /usr/local/lib/python3.6/dist-packages (from pandas->burst_detection) (2.8.1)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.6/dist-packages (from python-dateutil>=2.7.3->pandas->burst_detection) (1.15.0)\n",
            "Requirement already satisfied: torch in /usr/local/lib/python3.6/dist-packages (1.7.0+cu101)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.6/dist-packages (from torch) (3.7.4.3)\n",
            "Requirement already satisfied: future in /usr/local/lib/python3.6/dist-packages (from torch) (0.16.0)\n",
            "Requirement already satisfied: dataclasses in /usr/local/lib/python3.6/dist-packages (from torch) (0.7)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from torch) (1.18.5)\n",
            "Requirement already satisfied: torchvision in /usr/local/lib/python3.6/dist-packages (0.8.1+cu101)\n",
            "Requirement already satisfied: pillow>=4.1.1 in /usr/local/lib/python3.6/dist-packages (from torchvision) (7.0.0)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from torchvision) (1.18.5)\n",
            "Requirement already satisfied: torch==1.7.0 in /usr/local/lib/python3.6/dist-packages (from torchvision) (1.7.0+cu101)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.6/dist-packages (from torch==1.7.0->torchvision) (3.7.4.3)\n",
            "Requirement already satisfied: future in /usr/local/lib/python3.6/dist-packages (from torch==1.7.0->torchvision) (0.16.0)\n",
            "Requirement already satisfied: dataclasses in /usr/local/lib/python3.6/dist-packages (from torch==1.7.0->torchvision) (0.7)\n",
            "Reading package lists... Done\n",
            "Building dependency tree       \n",
            "Reading state information... Done\n",
            "subversion is already the newest version (1.9.7-4ubuntu1).\n",
            "0 upgraded, 0 newly installed, 0 to remove and 11 not upgraded.\n",
            "svn: E155000: Destination directory exists; please remove the directory or use --force to overwrite\n",
            "svn: E155000: 'datasets' already exists\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7BX-K0zdij37"
      },
      "source": [
        "## Objective\n",
        "\n",
        "* Understand anomalies in time-series\n",
        "* Ability to implement anomalies detection using forecasting with ARIMA, Kleinburg's Algorithm for Burst Detection, and Pytorch implementation of Autoencoder."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yv7wLhbVij38"
      },
      "source": [
        "import warnings\n",
        "warnings.filterwarnings('ignore')"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Wfdw1TD_ij4C",
        "outputId": "48f183c9-066f-4aae-989b-d5b5a3951820",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "# load packages\n",
        "\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "%pylab inline \n",
        "### Dataset to play with?\n",
        "from itertools import accumulate\n",
        "from plotnine import *\n",
        "import burst_detection as bd\n",
        "\n",
        "from sklearn import preprocessing\n",
        "\n",
        "import math\n",
        "import statsmodels.api as sm\n",
        "import statsmodels.tsa.api as smt\n",
        "from sklearn.metrics import mean_squared_error\n",
        "from matplotlib import pyplot\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Populating the interactive namespace from numpy and matplotlib\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kheMlFWaij4F"
      },
      "source": [
        "## Data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zTRUaCVhij4H"
      },
      "source": [
        "We use twitter data from the Numenta Anomaly Benchmark (NAB) v1.1. NAB is a novel benchmark for evaluating algorithms for anomaly detection in streaming, real-time applications. The twitter data contain number of tweets related to ten company ranging from Apple, Amazon to Google between the period of Feb 2015 to April 2015"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PhEa28bNij4I",
        "outputId": "f72e982c-085e-4f93-ee43-4d0c7fb10e3a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        }
      },
      "source": [
        "### Load All twitter data, identify where is the 'unusual traffic'\n",
        "tweets = ['AAPL','AMZN','CRM','CVS','FB','GOOG','IBM','KO','PFE','UPS']\n",
        "df = pd.read_csv('./datasets/Twitter_volume_AAPL.csv', parse_dates=True)\n",
        "df['key'] = 'AAPL'\n",
        "for i in tweets[1:]:\n",
        "    filename = './datasets/Twitter_volume_'+str(i)+'.csv'\n",
        "    file = pd.read_csv(filename, parse_dates=True)\n",
        "    file['key'] = i\n",
        "    df = pd.concat([df,file], sort=False)\n",
        "df"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>timestamp</th>\n",
              "      <th>value</th>\n",
              "      <th>key</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2015-02-26 21:42:53</td>\n",
              "      <td>104</td>\n",
              "      <td>AAPL</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2015-02-26 21:47:53</td>\n",
              "      <td>100</td>\n",
              "      <td>AAPL</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2015-02-26 21:52:53</td>\n",
              "      <td>99</td>\n",
              "      <td>AAPL</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2015-02-26 21:57:53</td>\n",
              "      <td>154</td>\n",
              "      <td>AAPL</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2015-02-26 22:02:53</td>\n",
              "      <td>120</td>\n",
              "      <td>AAPL</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15861</th>\n",
              "      <td>2015-04-22 23:27:53</td>\n",
              "      <td>3</td>\n",
              "      <td>UPS</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15862</th>\n",
              "      <td>2015-04-22 23:32:53</td>\n",
              "      <td>2</td>\n",
              "      <td>UPS</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15863</th>\n",
              "      <td>2015-04-22 23:37:53</td>\n",
              "      <td>2</td>\n",
              "      <td>UPS</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15864</th>\n",
              "      <td>2015-04-22 23:42:53</td>\n",
              "      <td>3</td>\n",
              "      <td>UPS</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15865</th>\n",
              "      <td>2015-04-22 23:47:53</td>\n",
              "      <td>3</td>\n",
              "      <td>UPS</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>158631 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                 timestamp  value   key\n",
              "0      2015-02-26 21:42:53    104  AAPL\n",
              "1      2015-02-26 21:47:53    100  AAPL\n",
              "2      2015-02-26 21:52:53     99  AAPL\n",
              "3      2015-02-26 21:57:53    154  AAPL\n",
              "4      2015-02-26 22:02:53    120  AAPL\n",
              "...                    ...    ...   ...\n",
              "15861  2015-04-22 23:27:53      3   UPS\n",
              "15862  2015-04-22 23:32:53      2   UPS\n",
              "15863  2015-04-22 23:37:53      2   UPS\n",
              "15864  2015-04-22 23:42:53      3   UPS\n",
              "15865  2015-04-22 23:47:53      3   UPS\n",
              "\n",
              "[158631 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U68SZ0TRij4L"
      },
      "source": [
        "The dataset is collected every five minutes, but we are not interested in that small of an inteval for time period so it is sum at hour level.\n",
        "\n",
        "Additionally, we also create a column for aggregate tweets for that company up until that moment and total number of tweets (for all 10 companies) for that hours. The reasoning for the total is to create a 'normal' behaviour for benchmarking"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yom39wfTij4L"
      },
      "source": [
        "df['timestamp'] =  pd.to_datetime(df['timestamp'], format='%Y-%m-%d %H:%M:%S')\n",
        "df['timestamp'] = df['timestamp'].apply(lambda dt: datetime.datetime(dt.year, dt.month, dt.day, dt.hour,dt.minute//60))\n",
        "df_alt = df.groupby(['timestamp'])['value'].agg('sum').reset_index(inplace = False)\n",
        "df = df.groupby(['timestamp','key'])['value'].agg('sum').reset_index(inplace = False)\n",
        "df = df.sort_values(by=['key','timestamp'])\n",
        "original = []\n",
        "for i in tweets:\n",
        "    output = [*accumulate(list(df[df['key']==i]['value']), lambda a, b: a+b)]\n",
        "    original= original + output\n",
        "df['aggregate'] = original\n",
        "df = df[df['value']!=0]\n",
        "df = df.merge(df_alt, on=['timestamp'], how='left')\n",
        "df.columns = ['timestamp','key','value','aggregate','total']"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gEjmDqyKij4O",
        "outputId": "1eb357f1-0d1c-4762-a3fd-d11821aea200",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 669
        }
      },
      "source": [
        "df.head(n=20)"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>timestamp</th>\n",
              "      <th>key</th>\n",
              "      <th>value</th>\n",
              "      <th>aggregate</th>\n",
              "      <th>total</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2015-02-26 21:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>457</td>\n",
              "      <td>457</td>\n",
              "      <td>1181</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2015-02-26 22:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1906</td>\n",
              "      <td>2363</td>\n",
              "      <td>3961</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2015-02-26 23:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>973</td>\n",
              "      <td>3336</td>\n",
              "      <td>2520</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2015-02-27 00:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>707</td>\n",
              "      <td>4043</td>\n",
              "      <td>2275</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2015-02-27 01:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>527</td>\n",
              "      <td>4570</td>\n",
              "      <td>1977</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>2015-02-27 02:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>782</td>\n",
              "      <td>5352</td>\n",
              "      <td>2752</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>2015-02-27 03:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>678</td>\n",
              "      <td>6030</td>\n",
              "      <td>3311</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>2015-02-27 04:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>492</td>\n",
              "      <td>6522</td>\n",
              "      <td>1692</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>2015-02-27 05:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>701</td>\n",
              "      <td>7223</td>\n",
              "      <td>2163</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>2015-02-27 06:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>393</td>\n",
              "      <td>7616</td>\n",
              "      <td>1411</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>2015-02-27 07:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>248</td>\n",
              "      <td>7864</td>\n",
              "      <td>1100</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>2015-02-27 08:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>255</td>\n",
              "      <td>8119</td>\n",
              "      <td>1108</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>2015-02-27 09:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>202</td>\n",
              "      <td>8321</td>\n",
              "      <td>1156</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>2015-02-27 10:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>283</td>\n",
              "      <td>8604</td>\n",
              "      <td>1423</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>2015-02-27 11:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>332</td>\n",
              "      <td>8936</td>\n",
              "      <td>1517</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>2015-02-27 12:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>404</td>\n",
              "      <td>9340</td>\n",
              "      <td>1598</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>2015-02-27 13:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>639</td>\n",
              "      <td>9979</td>\n",
              "      <td>2156</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>2015-02-27 14:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>680</td>\n",
              "      <td>10659</td>\n",
              "      <td>2495</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>2015-02-27 15:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>764</td>\n",
              "      <td>11423</td>\n",
              "      <td>2515</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>2015-02-27 16:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>898</td>\n",
              "      <td>12321</td>\n",
              "      <td>2838</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "             timestamp   key  value  aggregate  total\n",
              "0  2015-02-26 21:00:00  AAPL    457        457   1181\n",
              "1  2015-02-26 22:00:00  AAPL   1906       2363   3961\n",
              "2  2015-02-26 23:00:00  AAPL    973       3336   2520\n",
              "3  2015-02-27 00:00:00  AAPL    707       4043   2275\n",
              "4  2015-02-27 01:00:00  AAPL    527       4570   1977\n",
              "5  2015-02-27 02:00:00  AAPL    782       5352   2752\n",
              "6  2015-02-27 03:00:00  AAPL    678       6030   3311\n",
              "7  2015-02-27 04:00:00  AAPL    492       6522   1692\n",
              "8  2015-02-27 05:00:00  AAPL    701       7223   2163\n",
              "9  2015-02-27 06:00:00  AAPL    393       7616   1411\n",
              "10 2015-02-27 07:00:00  AAPL    248       7864   1100\n",
              "11 2015-02-27 08:00:00  AAPL    255       8119   1108\n",
              "12 2015-02-27 09:00:00  AAPL    202       8321   1156\n",
              "13 2015-02-27 10:00:00  AAPL    283       8604   1423\n",
              "14 2015-02-27 11:00:00  AAPL    332       8936   1517\n",
              "15 2015-02-27 12:00:00  AAPL    404       9340   1598\n",
              "16 2015-02-27 13:00:00  AAPL    639       9979   2156\n",
              "17 2015-02-27 14:00:00  AAPL    680      10659   2495\n",
              "18 2015-02-27 15:00:00  AAPL    764      11423   2515\n",
              "19 2015-02-27 16:00:00  AAPL    898      12321   2838"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cZKJqf2Fij4R"
      },
      "source": [
        "In NAB, we also have the answer. lA json file labelling all 'anomalies' datapoint which is given to us to help with validation."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4JRLuUoMij4R"
      },
      "source": [
        "### Read Answer Data\n",
        "import json\n",
        "# read file\n",
        "with open('./datasets/combined_labels.json', 'r') as myfile:\n",
        "    data=myfile.read()\n",
        "# parse file\n",
        "obj = json.loads(data)"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xR4C7jpoij4U"
      },
      "source": [
        "datetimenew = []\n",
        "keyword = []\n",
        "for i in tweets:\n",
        "    filename = 'realTweets/Twitter_volume_'+str(i)+'.csv'\n",
        "    size = len(obj[filename])\n",
        "    datetimenew += obj[filename]\n",
        "    keyword += [i]*size\n",
        "target = pd.DataFrame({'key':keyword,'timestamp':datetimenew,'outlier':1})\n",
        "target['timestamp'] =  pd.to_datetime(target['timestamp'], format='%Y-%m-%d %H:%M:%S')\n",
        "target['timestamp'] = target['timestamp'].apply(lambda dt: datetime.datetime(dt.year, dt.month, dt.day, dt.hour,dt.minute//60))"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kZnAyEa3ij4W",
        "outputId": "c70a5490-1609-40d6-bde6-2a2db15757e3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "target.head()"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>key</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>outlier</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>AAPL</td>\n",
              "      <td>2015-03-03 21:00:00</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>AAPL</td>\n",
              "      <td>2015-03-09 17:00:00</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>AAPL</td>\n",
              "      <td>2015-03-16 02:00:00</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>AAPL</td>\n",
              "      <td>2015-03-31 03:00:00</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>AMZN</td>\n",
              "      <td>2015-03-05 19:00:00</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    key           timestamp  outlier\n",
              "0  AAPL 2015-03-03 21:00:00        1\n",
              "1  AAPL 2015-03-09 17:00:00        1\n",
              "2  AAPL 2015-03-16 02:00:00        1\n",
              "3  AAPL 2015-03-31 03:00:00        1\n",
              "4  AMZN 2015-03-05 19:00:00        1"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DLvmKFa4ij4Y"
      },
      "source": [
        "df = df.merge(target, on=['key','timestamp'], how='left')\n",
        "df = df.fillna(0)"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tDRMXtMtVVPG"
      },
      "source": [
        "We now have a dataset with six columns containing everything we need to run some basic anomaly/outlier detection."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4oGqPxD2ij4b",
        "outputId": "b5d14a7c-fe1b-4754-805b-b04b7169fb3a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "df.head()"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>timestamp</th>\n",
              "      <th>key</th>\n",
              "      <th>value</th>\n",
              "      <th>aggregate</th>\n",
              "      <th>total</th>\n",
              "      <th>outlier</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2015-02-26 21:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>457</td>\n",
              "      <td>457</td>\n",
              "      <td>1181</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2015-02-26 22:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1906</td>\n",
              "      <td>2363</td>\n",
              "      <td>3961</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2015-02-26 23:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>973</td>\n",
              "      <td>3336</td>\n",
              "      <td>2520</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2015-02-27 00:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>707</td>\n",
              "      <td>4043</td>\n",
              "      <td>2275</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2015-02-27 01:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>527</td>\n",
              "      <td>4570</td>\n",
              "      <td>1977</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            timestamp   key  value  aggregate  total  outlier\n",
              "0 2015-02-26 21:00:00  AAPL    457        457   1181      0.0\n",
              "1 2015-02-26 22:00:00  AAPL   1906       2363   3961      0.0\n",
              "2 2015-02-26 23:00:00  AAPL    973       3336   2520      0.0\n",
              "3 2015-02-27 00:00:00  AAPL    707       4043   2275      0.0\n",
              "4 2015-02-27 01:00:00  AAPL    527       4570   1977      0.0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "usyCrWQYij4d"
      },
      "source": [
        "## Visualization"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e4xCEF5Wij4d"
      },
      "source": [
        "Before we get into the detail, let's see what is anomalies in NAB dataset and how does it look like. Some basic plot of each time-series and anomalies are shown below."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wb36C83Nij4e",
        "outputId": "d6acfd68-5519-459f-9498-6c621c398325",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        }
      },
      "source": [
        "ggplot(df, aes(x='timestamp',y='value', group='key', color='key'))+ geom_line()+ theme(\n",
        "  axis_text_x = element_text(size = 5.5,angle = 90, hjust = 20, vjust = -400)\n",
        ")+geom_point(df[df['outlier']==1],aes(x='timestamp',y='value', group='key'), color='black')+facet_wrap('~key', scales='free') "
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 640x480 with 10 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<ggplot: (8780128907286)>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OMWqtMYuij4h"
      },
      "source": [
        "What you may also noticed is that anomalies point that were given by NAB benchmark wasn't entirely accurate in all case or at the very least up for interpretation."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fcBiEkO3ij4h",
        "outputId": "0de5a23f-9344-45ee-a865-35437ae9e18e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "obj['realTweets/Twitter_volume_AAPL.csv']"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['2015-03-03 21:07:53',\n",
              " '2015-03-09 17:32:53',\n",
              " '2015-03-16 02:57:53',\n",
              " '2015-03-31 03:27:53']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9OgybX91ij4j",
        "outputId": "b6c2d307-bb15-4cf1-9ea3-6a5dcc772f19",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        }
      },
      "source": [
        "ggplot(df[df['key']=='AAPL'], aes(x='timestamp',y='value', group='key', color='key'))+ geom_line()+ theme(\n",
        "  axis_text_x = element_text(size = 5.5,angle = 90, hjust = 20, vjust = -400)\n",
        ")+geom_point(df[(df['outlier']==1)&(df['key']=='AAPL')],aes(x='timestamp',y='value', group='key'), color='black')\n"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<ggplot: (-9223363256725970295)>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LFlgsTTjV0zV"
      },
      "source": [
        "We also run visualization by 'aggregate data', this is less clear."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xo6Tbzbqij4m",
        "outputId": "791159f8-ee7a-4d30-b7a6-2f1664fa4288",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        }
      },
      "source": [
        "ggplot(df, aes(x='timestamp',y='aggregate', group='key', color='key'))+ geom_line()+ theme(\n",
        "  axis_text_x = element_text(size = 5.5,angle = 90, hjust = 20, vjust = -400)\n",
        ")+geom_point(df[df['outlier']==1],aes(x='timestamp',y='aggregate', group='key'), color='black')+facet_wrap('~key', scales='free') "
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 640x480 with 10 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<ggplot: (-9223363256725642331)>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2ysjIxJKij4o",
        "outputId": "ea4b1832-965f-46cd-bd70-c1ae37b53b27",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        }
      },
      "source": [
        "ggplot(df[df['key']=='AAPL'], aes(x='timestamp',y='aggregate', group='key', color='key'))+ geom_line()+ theme(\n",
        "  axis_text_x = element_text(size = 5.5,angle = 90, hjust = 20, vjust = -400)\n",
        ")+geom_point(df[(df['outlier']==1)&(df['key']=='AAPL')],aes(x='timestamp',y='aggregate', group='key'), color='black')"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<ggplot: (8780123212906)>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BUz3COmtij4r"
      },
      "source": [
        "## Anomalies"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LaSjXEfCij4r"
      },
      "source": [
        "After some basic visualization to understand what kind of anomalies existed in this example dataset., we now just need to give a brief explanation of an anomaly (or outlier). It is the deviation in a quantity from its expected value. Anomaly detection is the search of that which ‘does not conform’ with expected value."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "paDbGyNvij4s"
      },
      "source": [
        "* Point Anomaly\n",
        "![Point Anomaly](https://github.com/FreelancerFel/Anomalies_Detection/blob/master/global-outlier.png?raw=1)\n",
        "* Contextual Anomaly\n",
        "![Anomaly](https://github.com/FreelancerFel/Anomalies_Detection/blob/master/contextual-outlier.png?raw=1)\n",
        "* Collective Anomaly\n",
        "![Anomaly](https://github.com/FreelancerFel/Anomalies_Detection/blob/master/collective-outliers.png?raw=1)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rZbZ4WXJij4s"
      },
      "source": [
        "# Looking into the future - Using Forecasting to detect anomalies"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YRDENAUnij4s"
      },
      "source": [
        "One of the most simple way of detecting anomalies is to use time-series prediction. Essentially, anomalies is about finding datapoint that is out of ordinary. Time series forecasting helps us estimate that ordinary which we can then use to detect anomalies by comparing them with actual data.\n",
        "\n",
        "**It does requires you to set a threshold, but this can be navigate with multiple threshold if you are unsure.**\n",
        "\n",
        "This obviously relies on good forecasting model, but it also means that any forecasting model is compatible with this idea.\n",
        "\n",
        "* Exponential Smoothing\n",
        "* ARIMA\n",
        "* XGBoost - Lag Feature\n",
        "\n",
        "In this instance, we will simply use ARIMA which is essentially lag feature regression."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DViGlz1mij4t",
        "outputId": "31bfeafc-d9cd-4b81-90c0-e103196d2f88",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "AnomaliesData = df[df['key']=='AAPL']\n",
        "train, test = (AnomaliesData.value[0:-48]), (AnomaliesData.value[-48:])\n",
        "train_log, test_log = np.log10(train), np.log10(test) # take log to scale\n",
        "AnomaliesData.head()"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>timestamp</th>\n",
              "      <th>key</th>\n",
              "      <th>value</th>\n",
              "      <th>aggregate</th>\n",
              "      <th>total</th>\n",
              "      <th>outlier</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2015-02-26 21:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>457</td>\n",
              "      <td>457</td>\n",
              "      <td>1181</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2015-02-26 22:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1906</td>\n",
              "      <td>2363</td>\n",
              "      <td>3961</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2015-02-26 23:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>973</td>\n",
              "      <td>3336</td>\n",
              "      <td>2520</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2015-02-27 00:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>707</td>\n",
              "      <td>4043</td>\n",
              "      <td>2275</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2015-02-27 01:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>527</td>\n",
              "      <td>4570</td>\n",
              "      <td>1977</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            timestamp   key  value  aggregate  total  outlier\n",
              "0 2015-02-26 21:00:00  AAPL    457        457   1181      0.0\n",
              "1 2015-02-26 22:00:00  AAPL   1906       2363   3961      0.0\n",
              "2 2015-02-26 23:00:00  AAPL    973       3336   2520      0.0\n",
              "3 2015-02-27 00:00:00  AAPL    707       4043   2275      0.0\n",
              "4 2015-02-27 01:00:00  AAPL    527       4570   1977      0.0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6U2MdWS3ij4w"
      },
      "source": [
        "Determine P,D,Q through ACF and PACF table, alternatively use grid search according to either AIC, AICc or BIC value"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h8WtZfaeij4x",
        "outputId": "2bb449d3-af95-4903-ac48-278e5d135dd3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        }
      },
      "source": [
        "fig, axes = plt.subplots(1, 2, figsize=(15,4))\n",
        "fig = sm.graphics.tsa.plot_acf(train, lags=40, ax=axes[0])\n",
        "fig = sm.graphics.tsa.plot_pacf(train, lags=40, ax=axes[1])"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MDdmVVCCij4z",
        "outputId": "13fc8d33-2453-432c-d0cc-b62992007212",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "order_pdq = (1, 1, 1)\n",
        "seasonal_order_pdq = (0, 1, 1, 6)\n",
        "datapoint = [x for x in train_log]\n",
        "model = sm.tsa.SARIMAX(datapoint, order=order_pdq, seasonal_order=seasonal_order_pdq,enforce_stationarity=False,enforce_invertibility=False)\n",
        "res = model.fit(disp=False)\n",
        "print(res.summary())"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                                 Statespace Model Results                                \n",
            "=========================================================================================\n",
            "Dep. Variable:                                 y   No. Observations:                 1276\n",
            "Model:             SARIMAX(1, 1, 1)x(0, 1, 1, 6)   Log Likelihood                 181.208\n",
            "Date:                           Sun, 08 Nov 2020   AIC                           -354.416\n",
            "Time:                                   23:36:19   BIC                           -333.857\n",
            "Sample:                                        0   HQIC                          -346.691\n",
            "                                          - 1276                                         \n",
            "Covariance Type:                             opg                                         \n",
            "==============================================================================\n",
            "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
            "------------------------------------------------------------------------------\n",
            "ar.L1          0.7932      0.012     66.069      0.000       0.770       0.817\n",
            "ma.L1         -1.0000      2.761     -0.362      0.717      -6.412       4.412\n",
            "ma.S.L6       -1.0133      0.012    -87.482      0.000      -1.036      -0.991\n",
            "sigma2         0.0418      0.115      0.363      0.717      -0.184       0.268\n",
            "===================================================================================\n",
            "Ljung-Box (Q):                      133.56   Jarque-Bera (JB):              5984.54\n",
            "Prob(Q):                              0.00   Prob(JB):                         0.00\n",
            "Heteroskedasticity (H):               1.14   Skew:                             1.82\n",
            "Prob(H) (two-sided):                  0.18   Kurtosis:                        13.03\n",
            "===================================================================================\n",
            "\n",
            "Warnings:\n",
            "[1] Covariance matrix calculated using the outer product of gradients (complex-step).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C9vRbAcQij41"
      },
      "source": [
        "We are not trying to predict everything from scratch, but rather use information at each stage to predict what is likely to be the next stage. So instead of train/test split that are norm in forecasting, we reuse the data at every stage and make prediction only for the next datapoint."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F0o0FtAnij42",
        "outputId": "ecd6af07-95dc-441f-cf15-ff22dd60bf65",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "predictions = list()\n",
        "predict_log = list()\n",
        "for t in range(len(test_log)):\n",
        "    model = sm.tsa.SARIMAX(datapoint, order=order_pdq, seasonal_order=seasonal_order_pdq,enforce_stationarity=False,enforce_invertibility=False)\n",
        "    model_fit = model.fit(disp=0)\n",
        "    output = model_fit.forecast()\n",
        "    predict_log.append(output[0])\n",
        "    yhat = 10**output[0]\n",
        "    predictions.append(yhat)\n",
        "    obs = test_log.iloc[t]\n",
        "    datapoint.append(obs)\n",
        "predictions = pd.Series(predictions)\n",
        "predictions.index = test.index"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/statsmodels/base/model.py:512: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
            "  \"Check mle_retvals\", ConvergenceWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YP4_aIFdij44",
        "outputId": "36c4a62c-6461-4cb8-b1e8-9e69f1365e1b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 428
        }
      },
      "source": [
        "figsize=(15, 7)\n",
        "plt.figure(figsize=figsize)\n",
        "pyplot.plot(train,label='train', color='blue')\n",
        "pyplot.plot(test,label='test', color='purple')\n",
        "pyplot.plot(predictions, color='green',label='Predicted')\n",
        "pyplot.legend(loc='upper right')\n",
        "pyplot.show()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LGsD6yxFbski"
      },
      "source": [
        "Obviously, if you want to look at anomalies in a certain area, you shouldn't train your forecast model using that data. So we reserved some data at the end for some test."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1y9_gTMFij47",
        "outputId": "edd3f98a-6ad4-4591-9b41-63d60d1748dd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 428
        }
      },
      "source": [
        "figsize=(15, 7)\n",
        "plt.figure(figsize=figsize)\n",
        "pyplot.plot(test,label='test', color='purple')\n",
        "pyplot.plot(predictions, color='green',label='Predicted')\n",
        "pyplot.legend(loc='upper right')\n",
        "pyplot.show()"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3kAAAGbCAYAAAB541l1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeZiU1Zn38d/dS1VDdzU7dAMqCCgCbgiooxKXRNosEk2iMa+jjhqTjHEmk02dLI7J+L4xk4kZTTTDREezjCbjMppREdeQqCiIaIwN0uBGN0tDQ1c3TVf1ct4/6qm2gV5qeWrp6u/nurjsOvU8T53iD+THfZ9zzDknAAAAAEBhKMr1BAAAAAAA/iHkAQAAAEABIeQBAAAAQAEh5AEAAABAASHkAQAAAEABKcn1BFI1fvx4N23atFxPAwAAAABy4pVXXtnpnJtw4PiQDXnTpk3TmjVrcj0NAAAAAMgJM3u3r3HaNQEAAACggBDyAAAAAKCAEPIAAAAAoIAM2TV5AAAAAPJbR0eHtmzZovb29lxPZUgrKyvT1KlTVVpamtD1hDwAAAAAGbFlyxaFQiFNmzZNZpbr6QxJzjnt2rVLW7Zs0fTp0xO6h3ZNAAAAABnR3t6ucePGEfDSYGYaN25cUtVQQh4AAACAjCHgpS/Z30NCHgAAAAAUEEIeAAAAgIK0Z88e3X777Snd+5Of/ERtbW0+zyg7CHkAAAAACtJwDXnsrgkAAACgIF133XXatGmTjjvuOH3kIx/RxIkT9bvf/U6RSETnnXeebrzxRu3du1cXXHCBtmzZoq6uLn3nO9/R9u3b1dDQoDPOOEPjx4/Xs88+m+uvkhRCHgAAAICMW/6V5dq2bpuvz6w6rko1P6np9/0f/OAHeuONN7Ru3TqtWLFC999/v15++WU553Tuuedq5cqVamxs1OTJk/Xoo49KkpqbmzVq1Cj9+Mc/1rPPPqvx48f7OudsoF0TAAAAQMFbsWKFVqxYoeOPP17z58/X+vXrtXHjRh199NF68sknde211+qPf/yjRo0aleuppo1KHgAAAICMG6jilg3OOV1//fX6whe+cNB7a9eu1WOPPaZvf/vbOuuss/Td7343BzP0D5U8AMgzHfs61NnemetpAAAw5IVCIbW0tEiSlixZorvuukutra2SpPr6eu3YsUMNDQ0aOXKkLr74Yn3jG9/Q2rVrD7p3qKGSBwB55v4L7lfZmDKd98vzcj0VAACGtHHjxumUU07RvHnzdM455+hzn/ucTj75ZElSRUWFfv3rX6uurk7f+MY3VFRUpNLSUt1xxx2SpKuuuko1NTWaPHnykNt4xZxzuZ5DShYsWODWrFmT62kAgO9+NudnGjFmhC5//vJcTwUAgLTU1tbqqKOOyvU0CkJfv5dm9opzbsGB11LJA4A8EwlHVFRCNz0AAEgNIQ8A8kwkHFFxaXGupwEAAIYoQh4A5BHX7RRtiRLyAABAyugHAoA8Em2NSpIiLREN1TXTAAAgtwh5AJBHIuGIJKm7o1tdka4czwYAAAxFhDwAyCPxkCfFqnkAAADJIuQBQB7pHfKiLdEczgQAgMJQXFys4447TvPmzdNnPvMZtbW1pfysyy67TPfff78k6corr9Sbb77Z77XPPfecXnjhhaQ/Y9q0adq5c2fKc5QSCHlmdpeZ7TCzNw4Yv8bM1pvZX8zsh73GrzezOjPbYGZLeo3XeGN1ZnZdr/HpZvaSN/5bMwuk9Y0AYAjbr5IXppIHAEC6RowYoXXr1umNN95QIBDQz3/+8/3e7+zsTOm5v/jFLzRnzpx+30815PkhkUre3ZJqeg+Y2RmSlko61jk3V9KPvPE5kj4raa53z+1mVmxmxZJ+JukcSXMkXeRdK0k3S7rFOTdT0m5JV6T7pQBgqKJdEwCAzDnttNNUV1en5557TqeddprOPfdczZkzR11dXfrGN76hhQsX6phjjtG///u/S5Kcc/ryl7+sI488Uh/+8Ie1Y8eOnmedfvrpWrNmjSRp+fLlmj9/vo499lidddZZeuedd/Tzn/9ct9xyi4477jj98Y9/VGNjoz71qU9p4cKFWrhwoZ5//nlJ0q5du3T22Wdr7ty5uvLKK33ZeG3QIxSccyvNbNoBw1+S9APnXMS7Jv5tl0q6zxt/28zqJC3y3qtzzm2WJDO7T9JSM6uVdKakz3nX3CPpnyTdkeoXAoChjHZNAECh+sryr2jdtnW+PvO4quP0k5qfJHRtZ2enHn/8cdXUxOpXa9eu1RtvvKHp06dr2bJlGjVqlFavXq1IJKJTTjlFZ599tl599VVt2LBBb775prZv3645c+bo8ssv3++5jY2N+vznP6+VK1dq+vTpampq0tixY/XFL35RFRUV+vrXvy5J+tznPqd/+Id/0Kmnnqr33ntPS5YsUW1trW688Uadeuqp+u53v6tHH31Ud955Z9q/L6mek3eEpNPM7CZJ7ZK+7pxbLWmKpFW9rtvijUnS+weMnyhpnKQ9zrnOPq4/iJldJekqSTr00ENTnDoA5C/aNQEA8Ne+fft03HHHSYpV8q644gq98MILWrRokaZPny5JWrFihV5//fWe9XbNzc3auHGjVq5cqYsuukjFxcWaPHmyzjzzzIOev2rVKi1evLjnWWPHju1zHk899dR+a/jC4bBaW1u1cuVKPfjgg5Kkj33sYxozZkza3znVkFciaaykkyQtlPQ7Mzs87dkMwjm3TNIySVqwYAEHSAEoOLRrAgAKVaIVN7/F1+QdqLy8vOdn55xuu+02LVmyZL9rHnvsMd/m0d3drVWrVqmsrMy3Z/Yn1d01t0h60MW8LKlb0nhJ9ZIO6XXdVG+sv/FdkkabWckB4wAwLEXCEcl6/QwAADJuyZIluuOOO9TR0SFJeuutt7R3714tXrxYv/3tb9XV1aWtW7fq2WefPejek046SStXrtTbb78tSWpqapIkhUIhtbS09Fx39tln67bbbut5HQ+eixcv1n/9139Jkh5//HHt3r077e+Tasj7H0lnSJKZHSEpIGmnpEckfdbMgmY2XdIsSS9LWi1plreTZkCxzVkecbFVhc9K+rT33EslPZzqlwGAoa69uV3lE2L/ssiaPAAAsuPKK6/UnDlzNH/+fM2bN09f+MIX1NnZqfPOO0+zZs3SnDlzdMkll+jkk08+6N4JEyZo2bJlOv/883XsscfqwgsvlCR94hOf0EMPPdSz8cqtt96qNWvW6JhjjtGcOXN6dvm84YYbtHLlSs2dO1cPPvigL8vSbLDdW8zsXkmnK1ap2y7pBkm/knSXpOMkRRVbk/eMd/23JF0uqVPSV5xzj3vjH5X0E0nFku5yzt3kjR8u6T7F2j9flXRxfEOXgSxYsMDFd7MBgEJx/4X3a9tr2xTeEtYJXzhBS/51yeA3AQCQp2pra3XUUUflehoFoa/fSzN7xTm34MBrE9ld86J+3rq4n+tvknRTH+OPSTqoqdXbcXPRgeMAMBxFwhEFK4MKhoK0awIAgJSk2q4JAMiAeMgLhAK0awIAgJQQ8gAgj/RU8iqp5AEACoMfh3sPd8n+HhLyACCP9G7XpJIHABjqysrKtGvXLoJeGpxz2rVrV1JHL6R6Th4AIAN6V/L2vLsn19MBACAtU6dO1ZYtW9TY2JjrqQxpZWVlmjp1asLXE/IAIE845/Zbk0e7JgBgqCstLdX06dNzPY1hh3ZNAMgTHW0dct2OjVcAAEBaCHkAkCfilbuejVdaqOQBAIDkEfIAIE/sF/JCQXVFutQV7crxrAAAwFBDyAOAPNE75AVCgdgY1TwAAJAkQh4A5IkD2zV7jwEAACSKkAcAeeLAdk1JbL4CAACSRsgDgDzRE/JG9ark0a4JAACSRMgDgDzR55o82jUBAECSCHkAkCd6Ql6Idk0AAJA6Qh4A5IlIOKKSshIVB4rZeAUAAKSMkAcAeSISjvSEO45QAAAAqSLkAUCeiIajPSGPdk0AAJAqQh4A5InelbyikiKVjCihXRMAACSNkAcAeaJ3yJNi1TzaNQEAQLIIeQCQJw4KeZVB2jUBAEDSCHkAkCcODHmBUIB2TQAAkDRCHgDkiUg4okBloOd1MEQlDwAAJI+QBwB5oq92TSp5AAAgWYQ8AMgDnZFOdUW7Dm7XZOMVAACQJEIeAOSBeMWOjVcAAEC6CHkAkAf6CnlsvAIAAFJByAOAPNBnJS8UVGd7p7o7u3M1LQAAMAQR8gAgD0Sa+27XlMS6PAAAkBRCHgDkgf7aNXu/BwAAkAhCHgDkgf42XpHE5isAACAphDwAyAP9rcnr/R4AAEAiCHkAkAcGbNdkTR4AAEgCIQ8A8kAkHFFRSZFKykp6xmjXBAAAqRg05JnZXWa2w8ze6OO9r5mZM7Px3mszs1vNrM7MXjez+b2uvdTMNnq/Lu01foKZ/dm751YzM7++HAAMFZFwRMHKoHr/EUi7JgAASEUilby7JdUcOGhmh0g6W9J7vYbPkTTL+3WVpDu8a8dKukHSiZIWSbrBzMZ499wh6fO97jvoswCg0MVDXm8coQAAAFIxaMhzzq2U1NTHW7dI+qYk12tsqaRfuphVkkabWbWkJZKedM41Oed2S3pSUo33XqVzbpVzzkn6paRPpveVAGDo6SvkcYQCAABIRUpr8sxsqaR659xrB7w1RdL7vV5v8cYGGt/Sx3h/n3uVma0xszWNjY2pTB0A8lJfIa+4tFjFwWLW5AEAgKQkHfLMbKSkf5T0Xf+nMzDn3DLn3ALn3IIJEyZk++MBIGP6CnlSrGWTdk0AAJCMVCp5MyRNl/Samb0jaaqktWZWJale0iG9rp3qjQ00PrWPcQAYVvoNeaGgomEqeQAAIHFJhzzn3J+dcxOdc9Occ9MUa7Gc75zbJukRSZd4u2yeJKnZObdV0hOSzjazMd6GK2dLesJ7L2xmJ3m7al4i6WGfvhsADBmRcESBysBB44FQgEoeAABISiJHKNwr6UVJR5rZFjO7YoDLH5O0WVKdpP+Q9LeS5JxrkvR9Sau9X9/zxuRd8wvvnk2SHk/tqwDA0DVQuyZr8gAAQDJKBrvAOXfRIO9P6/Wzk3R1P9fdJemuPsbXSJo32DwAoFB1dXSpc19nv+2ardtaczArAAAwVKW0uyYAwD/xSh0brwAAAD8Q8gAgx+Ln4JWNKjvovUAowDl5AAAgKYQ8AMixeIjrq5IXCAVYkwcAAJJCyAOAHBso5AUrg+po61B3V3e2pwUAAIYoQh4A5NiAIS8UG6OaBwAAEkXIA4AcG6ySJ4nNVwAAQMIIeQCQY4Otyet9DQAAwGAIeQCQY7RrAgAAPxHyACDHIuGIZFJpeelB79GuCQAAkkXIA4Aci4QjClYGZWYHvUe7JgAASBYhDwByLB7y+kK7JgAASBYhDwBybMCQR7smAABIEiEPAHJsoJBHuyYAAEgWIQ8AcmygkFcSLFFxoJh2TQAAkDBCHgDk2EAhT4pV86jkAQCARBHyACDHBgt5wVCQSh4AAEgYIQ8AcizSPEjIqwyy8QoAAEgYIQ8Acqi7q1vR1ijtmgAAwDeEPADIoWhrrA1zsEoe7ZoAACBRhDwAyKF4hW6wNXm0awIAgEQR8gAghxIJebRrAgCAZBDyACCHEqrk0a4JAACSQMgDgBxKtJIXbY3KdbtsTQsAAAxhhDwAyKFE1+RJH2zSAgAAMBBCHgDkUKLtmpLYfAUAACSEkAcAOZRou2bvawEAAAZCyAOAHIoHt3iQ60s8ALL5CgAASAQhDwByKBKOqLS8VEXF/f9xHF+TRyUPAAAkgpAHADkUCUcGbNWUerVrsiYPAAAkgJAHADkUDUcHDXm0awIAgGQQ8gAghxKp5NGuCQAAkkHIA4AcSijkcYQCAABIAiEPAHIokZBXHCxWUUkR7ZoAACAhhDwAyKFIOKKyUWUDXmNmCoQCtGsCAICEDBryzOwuM9thZm/0GvsXM1tvZq+b2UNmNrrXe9ebWZ2ZbTCzJb3Ga7yxOjO7rtf4dDN7yRv/rZn1f1gUABSYSDiiQOXgf+wFK4NU8gAAQEISqeTdLanmgLEnJc1zzh0j6S1J10uSmc2R9FlJc717bjezYjMrlvQzSedImiPpIu9aSbpZ0i3OuZmSdku6Iq1vBABDhHMuoXZNKbb5CpU8AACQiEFDnnNupaSmA8ZWOOc6vZerJE31fl4q6T7nXMQ597akOkmLvF91zrnNzrmopPskLTUzk3SmpPu9+++R9Mk0vxMADAkdbR1y3S6xkFcZZOMVAACQED/W5F0u6XHv5ymS3u/13hZvrL/xcZL29AqM8fE+mdlVZrbGzNY0Njb6MHUAyJ14ZS6RkBcIBWjXBAAACUkr5JnZtyR1SvqNP9MZmHNumXNugXNuwYQJE7LxkQCQMcmEPNo1AQBAokpSvdHMLpP0cUlnOeecN1wv6ZBel031xtTP+C5Jo82sxKvm9b4eAApaUpW8ygDtmgAAICEpVfLMrEbSNyWd65xr6/XWI5I+a2ZBM5suaZaklyWtljTL20kzoNjmLI944fBZSZ/27r9U0sOpfRUAGFqo5AEAgExI5AiFeyW9KOlIM9tiZldI+qmkkKQnzWydmf1ckpxzf5H0O0lvSlou6WrnXJdXpfuypCck1Ur6nXetJF0r6atmVqfYGr07ff2GAJCnkl6T1xrVB40TAAAAfRu0XdM5d1Efw/0GMefcTZJu6mP8MUmP9TG+WbHdNwFgWEmqklcZlJzUsbdDgQqOEwUAAP3zY3dNAEAKkm3X7H0PAABAfwh5AJAjPSEvlGAlT2LzFQAAMChCHgDkSCQcUUlZiYoDxYNeGwjFWjQ5Kw8AAAyGkAcAORJpjiTUqinRrgkAABJHyAOAHImEkwh5tGsCAIAEEfIAIEeSCXnxdk0qeQAAYDCEPADIkVQqeazJAwAAgyHkAUCOJBXyQrRrAgCAxBDyACBHkgl5JSNKZEVGuyYAABgUIQ8AciQSjihQGUjoWjNTsDJIuyYAABgUIQ8AcsA5l1QlT4ptvkLIAwAAgyHkAUAOdEW61N3RnVTIC4aCtGsCAIBBEfIAIAfiYS2pkFcZZOMVAAAwKEIeAORAKiEvEApQyQMAAIMi5AFADqRayWNNHgAAGAwhDwByIKWQF6JdEwAADI6QBwA5QLsmAADIFEIeAORAOu2azrlMTQsAABQAQh4A5ECqlTzX7dTR1pGpaQEAgAJAyAOAHEi1kieJzVcAAMCACHkAkAORcERFJUUqKStJ+J5gKBby2HwFAAAMhJAHADkQCUcUrAzKzBK+JxAK9NwLAADQH0IeAORAJBxRcFTirZoS7ZoAACAxhDwAyIF4JS8ZtGsCAIBEEPIAIAdSCnne9bRrAgCAgRDyACAHUgl58TV5tGsCAICBEPIAIAfSatekkgcAAAZAyAOAHEgl5JWWl0rGmjwAADAwQh4A5EAqIc/MFAwFadcEAAADIuQBQJZ1dXSpc19n0iFPiq3Lo10TAAAMhJAHAFkWr8SlEvKClVTyAADAwAh5AJBl8UpcSiEvFGRNHgAAGBAhDwCyLK2QVxmkXRMAAAxo0JBnZneZ2Q4ze6PX2Fgze9LMNnr/HeONm5ndamZ1Zva6mc3vdc+l3vUbzezSXuMnmNmfvXtuNTPz+0sCQD5JJ+QFQgHaNQEAwIASqeTdLanmgLHrJD3tnJsl6WnvtSSdI2mW9+sqSXdIsVAo6QZJJ0paJOmGeDD0rvl8r/sO/CwAKCjtze2S0mjXpJIHAAAGMGjIc86tlNR0wPBSSfd4P98j6ZO9xn/pYlZJGm1m1ZKWSHrSOdfknNst6UlJNd57lc65Vc45J+mXvZ4FAAUprUpeZYA1eQAAYECprsmb5Jzb6v28TdIk7+cpkt7vdd0Wb2yg8S19jPfJzK4yszVmtqaxsTHFqQNAbqW78Uq0JarYv4sBAAAcLO2NV7wKXFb+tuGcW+acW+CcWzBhwoRsfCQA+C7djVe6O7vV2d7p97QAAECBSDXkbfdaLeX9d4c3Xi/pkF7XTfXGBhqf2sc4ABSsSDgimVRaXpr0vYFQQJLYfAUAAPQr1ZD3iKT4DpmXSnq41/gl3i6bJ0lq9to6n5B0tpmN8TZcOVvSE957YTM7ydtV85JezwKAghQJRxSsDCqVzYSDoWDPMwAAAPpSMtgFZnavpNMljTezLYrtkvkDSb8zsyskvSvpAu/yxyR9VFKdpDZJfyNJzrkmM/u+pNXedd9zzsU3c/lbxXbwHCHpce8XABSsaDiaUqum9EGLJ5uvAACA/gwa8pxzF/Xz1ll9XOskXd3Pc+6SdFcf42skzRtsHgBQKOKVvFTQrgkAAAaT9sYrAIDkpBPyaNcEAACDIeQBQJalFfJo1wQAAIMg5AFAltGuCQAAMomQBwBZ5kslj3ZNAADQD0IeAGRZWpW88lglb6B2zfpwvV6ufzml5wMAgKGPkAcAWdTd1a1oa+pHKFiRKVARGLCSd+MfbtQn7v1EqlMEAABDHCEPALIo2hpbS5dqyIvfO9CavHeb39WOvTsU6aSlEwCA4YiQBwBZFK/ApRPyAqHAgCGvPlwvSdq+d3vKnwEAAIYuQh4AZJEfIS9YGRywXbO+JRbytrVuS/kzAADA0EXIA4As8iXkhYL9brzS1tGmPe17JElbW7am/BkAAGDoIuQBQBb51a7ZXyWvoaWh52cqeQAADE+EPADIIr/aNftbkxdfjydJW1up5AEAMBwR8gAgi3pC3qg0K3n9tGtSyQMAAIQ8AMgi39bk9dOuGd905bBRh1HJAwBgmCLkAUAWxcNZoCKQ8jOClUF1d3SrM9J50Hv14XqVl5bryPFHUskDAGCYIuQBQBZFwhEFKgIqKk79j99AKBYQ+1qXV99SrymVU1RdUc3umgAADFOEPADIokg4klarpvRBq2dfLZsNLQ2aHJqsqooqbWvdJudcWp8FAACGHkIeAGRRNBxNP+SFvJDXx+Yr9S31mhKKVfI6ujvUtK8prc8CAABDDyEPALLIj0pevF3zwEqec04NLQ2aEpqiqooqSeywCQDAcETIA4As8rNd88A1ebv27VK0K6rJocmqDlVL4qw8AACGI0IeAGSRLyGvn3bN+EHoUyqp5AEAMJwR8gAgizK58Ur8jLz4mjxJ7LAJAMAwRMgDgCxqb25XoDL1M/Kk/o9QaGhpkCRNDk1WRaBCI0tH0q4JAMAwRMgDgCxxzvmz8Yp3kHp/7ZrVoWqZmaorqmnXBABgGCLkAUCWdOztkJzSDnlFxUUqLS/ts11zYvlEBYpjIbA6VE0lDwCAYYiQBwBZEg9l6YY8Kbb5yoHtmvEz8uLiB6IDAIDhhZAHAFnia8irDB5UyWtoadDk0OSe19UV1Wy8AgDAMETIA4As8TPkBUKBgyt54YMrec2RZu3r2Jf25wEAgKGDkAcAWeJ3u2bvjVcinRE1tjVqSuUHIS9+jAItmwAADC+EPADIkky2a8aDXO92TQ5EBwBgeCLkAUCWZLJds/dB6HHVIe9AdHbYBABgWCHkAUCW+B3yelfy4mfk9W7XpJIHAMDwRMgDgCzpCXkhn9o1e63Ji1fyerdrThg5QUVWxA6bAAAMM2mFPDP7BzP7i5m9YWb3mlmZmU03s5fMrM7MfmtmAe/aoPe6znt/Wq/nXO+NbzCzJel9JQDIT5FwRCVlJSoOFKf9rGAoqK5Il7qiXZJixycEi4MaN2JczzXFRcWaWD6RSh4AAMNMyiHPzKZI+jtJC5xz8yQVS/qspJsl3eKcmylpt6QrvFuukLTbG7/Fu05mNse7b66kGkm3m1n6fwMCgDwTCUd8adWUPmj5jFfz6lvqNTk0WWa233XVFdWsyQMAYJhJt12zRNIIMyuRNFLSVklnSrrfe/8eSZ/0fl7qvZb3/lkW+9vIUkn3Oecizrm3JdVJWpTmvAAg7/gZ8gKhgCT1bL5SH67fbz1eXFVFFZU8AACGmZRDnnOuXtKPJL2nWLhrlvSKpD3OuU7vsi2S4n/rmCLpfe/eTu/6cb3H+7hnP2Z2lZmtMbM1jY2NqU4dAHLC10peaP9KXkNLw37r8eKo5AEAMPyk0645RrEq3HRJkyWVK9ZumTHOuWXOuQXOuQUTJkzI5EcBgO8y0q4Zjsg5p/qW+v2OT4irqqjS9tbt6nbdvnwuAADIf+m0a35Y0tvOuUbnXIekByWdImm0174pSVMl1Xs/10s6RJK890dJ2tV7vI97AKBgZKpdsznSrLaOtj5DXnWoWl2uSzvbdvryuQAAIP+lE/Lek3SSmY301tadJelNSc9K+rR3zaWSHvZ+fsR7Le/9Z5xzzhv/rLf75nRJsyS9nMa8ACAvZaqS19DSIEl9tmtyVh4AAMNPOmvyXlJsA5W1kv7sPWuZpGslfdXM6hRbc3end8udksZ541+VdJ33nL9I+p1iAXG5pKudc12pzgsA8lUkHFGgMuDLs3qvyevrIPS46opqSeKsPAAAhpGSwS/pn3PuBkk3HDC8WX3sjumca5f0mX6ec5Okm9KZCwDkM+dcxto14weh97cmT6KSBwDAcJLuEQoAgAR0RbrU3dHt/+6a4Q8qeQO1a7LDJgAAwwchDwCyIBKOHXXgV8grKilSyYgSRVpia/LGlI3RiNIRB11XHihXKBCikgcAwDBCyAOALIiHvLJRZb49MxgKxip5LX0fhB5XHeKsPAAAhhNCHgBkgd+VvPiz4mvy+lqPF1dVUUUlDwCAYYSQBwBZkImQFwgFFG2JqqGloc/1eHHVFdXsrgkAwDBCyAOALMhUJa+tpU3bWrdRyQMAAD0IeQCQBRkJeaGgGqON6nbdA6/Jq6hWS7RFe6N7fftsAACQvwh5AJAFmWrX3Nm9U1LfxyfEcVYeAADDCyEPALIgU+2aOxULeQO1a1aHqiVxVh4AAMMFIQ8AsiASjqiotEjFwWLfnhkIBbSreJckDdiuSSUPAIDhhZAHAFnQ3tyuYGVQZubbM4OVQe0ZsUfFVqyJ5RP7va66wqvkscMmAADDAiEPALIgGo762qopxTZeaQm1qKq8SkXW/x/n40aOU0lRCZU8AACGCUIeAGRBJBzxPeQFQgG1hFpUPaJ6wOuKrEiTyiexJg8AgCD6y0gAACAASURBVGGCkAcAWZCJkBesDCpcGdak0kmDXstZeQAADB+EPADIgoyEPK9dc0LxhEGvrQ5VU8kDAGCYIOQBQBZkIuR1juxUpCyi8TZ+0GuryqvYeAUAgGGCkAcAWZCJkNdU0iRJGtc1btBrq0PVamxrVFd3l69zAAAA+YeQBwBZkImQt8tiZ+SN7hg96LXVFdXqdt3asXeHr3MAAAD5h5AHABnWFe1SZ3un7yFvp3ZKkkbvGzzkcSA6AADDByEPADIs0hKRJN9D3vaO7ZKk0XsTqOSFvAPR2XwFAICCV5LrCQBAoYuEMxPytu7dqmAkqOKu4kGvpZIHAMDwQcgDgAzLVMhraG1QZVulIpHIoNfGQx47bAIAUPgIeQCQYZkKefXheo1uH61oa3TQa8tKyjS6bDSVPAAAhgHW5AFAhmUs5LXUa2zH2J7nD6a6ggPRAQAYDgh5AJBhmQh53a5bW1u2amz3WEVbBq/kSbGWTSp5AAAUPkIeAGRYJkLezrad6uju0ASb0LN752CqQ1TyAAAYDgh5AJBhmQh59eF6SdLE4okJt2tWlccqec453+YBAADyDyEPADIsEo5IJpWWl/r2zPoWL+QFJibcrlkdqlZbR5taoi2+zQMAAOQfQh4AZFgkHFGwMigz8+2ZDS0NkqTqkdUJt2tyVh4AAMMDIQ8AMiwajmbk+ASTqaqiSh17O9Td1T3oPdUV1ZI4Kw8AgEJHyAOADItX8vxU31KvSRWTVB4ql6SEzsqjkgcAwPBAyAOADMtEyGtoadDk0OSe5yay+Up1yKvkscMmAAAFjZAHABmWqUrelNAUBUIBSUpo85UxZWMUKA5QyQMAoMAR8gAgwzIS8sKxkBcMeZW8BDZfMYut4aOSBwBAYUsr5JnZaDO738zWm1mtmZ1sZmPN7Ekz2+j9d4x3rZnZrWZWZ2avm9n8Xs+51Lt+o5ldmu6XAoB84nfIi3RGtGvfrqTbNaXYujwqeQAAFLZ0K3n/Jmm5c262pGMl1Uq6TtLTzrlZkp72XkvSOZJmeb+uknSHJJnZWEk3SDpR0iJJN8SDIQAUgkg4ouAo/0Je/PiEKZXJtWtKsR022V0TAIDClnLIM7NRkhZLulOSnHNR59weSUsl3eNddo+kT3o/L5X0SxezStJoM6uWtETSk865JufcbklPSqpJdV4AkE+6u7oVbfX3CIX4QehTQlOo5AEAgIOkU8mbLqlR0n+a2atm9gszK5c0yTkX/2fibZImeT9PkfR+r/u3eGP9jR/EzK4yszVmtqaxsTGNqQNAdsSPNvA15IVjIW9yaHJSa/KkWCWvsa1RHV0dvs0HAADkl3RCXomk+ZLucM4dL2mvPmjNlCQ555wkl8Zn7Mc5t8w5t8A5t2DChAl+PRYAMiZeYfMz5KXTrhk/K2/H3h2+zQcAAOSXdELeFklbnHMvea/vVyz0bffaMOX9N/43iXpJh/S6f6o31t84AAx5mQh59S31Kisp05iyMSoJlqg4UJxwuyZn5QEAUPhSDnnOuW2S3jezI72hsyS9KekRSfEdMi+V9LD38yOSLvF22TxJUrPX1vmEpLPNbIy34crZ3hgADHmZCnlTQlNkZpKkQCiQcLtmvJLHujwAAApXSZr3XyPpN2YWkLRZ0t8oFhx/Z2ZXSHpX0gXetY9J+qikOklt3rVyzjWZ2fclrfau+55zrinNeQFAXog0Z6Zdc3Jocs/rYGUwqd01JbHDJgAABSytkOecWydpQR9vndXHtU7S1f085y5Jd6UzFwDIRxmp5IXrtXDKwp7XwVAw4XbNSRWxvbCo5AEAULjSPScPADAAv0Oec66nXTMuEAokXMkLFAc0bsQ41uQBAFDACHkAkEF+h7w97XvU3tl+ULtmopU8ibPyAAAodIQ8AMigePgKVAR8eV7vg9DjgqFgwhuvSLEdNqnkAQBQuAh5AJBBkXBEgYqAior9+eM2fhD6lMrU2jUlKnkAABQ6Qh4AZFAkHPH9+ARJabVrVldUa2vLVsX2wwIAAIWGkAcAGeR3yGtoaZC0f8gLhAKKtkbluhMLbVUVVYp0RdQcafZtXgAAIH8Q8gAgg3yv5IXrNW7EOJWVlPWMxZ8f3ctZeQAAgJAHABmViXbN3uvxpNjGK/HPSkR1yAt5bL4CAEBBIuQBQAZlol2zd6umFGvXlJTw5itVFVWSOBAdAIBCRcgDgAzKSCUvdEAlrzLJSh7tmgAAFDRCHgBkUCQcUaDSnzPyOro6tL11+8EhL96umeBZeZXBSpWVlFHJAwCgQBHyACBDnHO+VvK2tW6TkzuoXbNn45UE2zXNLHaMAmvyAAAoSIQ8AMiQjr0dkpNvIS9+fMKBG6/E1+Qlc1YeB6IDAFC4CHkAkCHx0OVXyIsfhJ5uu6YU22GTSh4AAIWJkAcAGeJ7yAt7Ie/AIxSS3HhFkqrKqeQBAFCoCHkAkCF+h7yGlgaVFpVq/Mjx+40XB4tVVFKU8Jo8KVbJa9rXpEhn4sEQAAAMDYQ8AMiQTLRrVoeqVWT7/9FtZgqEAkm1a8bPytu+d7svcwMAAPmDkAcAGZKJkHfgery4YGVQ0XASlTzOygMAoGAR8gAgQzLRrnng8QlxwVAwpUoe6/IAACg8hDwAeeX5f3let868Vd1d3bmeStoysfHKgJW8JNfkSWKHTQAAChAhD0BeWXfXOu3etFsNqxtyPZW09YS8UPohryXSopZoy0E7a8YFQoGkdtecWD5RJqOSBwBAASLkAcgbjW82auf6nZKkuifqcjyb9EXCEZWUlag4UJz2s+Jn5PnVrllSVKIJ5RNYkwcAQAEi5AHIG7UP1kqSxs4cq03LN+V4NumLhCO+rseTDj4IPS5QmVwlT4qty9u2l0oeAACFhpAHIG/UPlCrQ/7qEB39f45W/cv1atvVlusppcXPkNffQehxwVBya/Kk2A6bVPIAACg8hDwAeWH35t3atm6bZp8/WzOWzJDrdtr81OZcTystkXBEwVH+HZ8gDdCuWRlr13TOJfzMqooq1uQBAFCACHkA8kLtQ7FWzaPOP0pTFk5R2ZiyId+y6Xe7ZmWwUhWBij7fD4QCkpM69nYk/Mzqimpta92WVDAEAAD5j5AHIC/UPlCrquOrNGb6GBWVFGnGR2aobnndkA4gvrZrDnAQuvTBDp7JnpXX0d2hpn1Nac8PAADkD0IegJxraWjRlhe36KhPHdUzNqNmhlq3tWr769tzOLP0+L0mr7/1eNIHZ/Els/kKZ+UBAFCYCHkAcq53q2bczCUzJUl1y4fuUQqRZn/bNftbjyd57ZpSUpuvVFVUSRLr8gAAKDCEPAA5t/7B9Rp/1HhNOGpCz1hockiTjpk0ZNflOed8q+R1u25tbd06cLtmZfLtmtUVXiWPHTYBACgohDwAOdW2s03v/OGd/ap4cTNqZui9P72XVHDJF53tneru7PYl5O3Yu0Od3Z2JrclLol2TSh4AAIWJkAcgpzY8skGuy/UZ8mbWzFR3Z7feefad7E8sTfGw5UfIi5+R53e7ZigYUnlpOWvyAAAoMIQ8ADlV+2CtRk8brarjqw5679BTDlVpeemQXJfnZ8hraGmQ1P9B6L0/J9mqJ2flAQBQeNIOeWZWbGavmtn/eq+nm9lLZlZnZr81s4A3HvRe13nvT+v1jOu98Q1mtiTdOQEYGiLhiDY/uVmzz58tMzvo/eJAsaafOV11jw+9oxR8reR5B6H73a4pxXbYpJIHAEBh8aOS9/eSanu9vlnSLc65mZJ2S7rCG79C0m5v/BbvOpnZHEmflTRXUo2k282s2Id5Achzbz36lrqiXZrzqTn9XjOzZqb2vLNHTRuH1llufrdrFlmRJlVM6veakhElsiJLql1TopIHAEAhSivkmdlUSR+T9AvvtUk6U9L93iX3SPqk9/NS77W898/yrl8q6T7nXMQ597akOkmL0pkXgKFh/YPrVVFVoaknTe33mpk1Q/MoBb/bNSeVT1JJUUm/15iZgpXB5Ct5FdXsrgkAQIFJt5L3E0nflNTtvR4naY9zrtN7vUVSvL9oiqT3Jcl7v9m7vme8j3v2Y2ZXmdkaM1vT2NiY5tQB5FJHW4c2PrZRs8+bLSs6uFUzbszhYzR21thhHfLqWwY+CD0uEAqkVMlrjjRrX8e+VKcHAADyTMohz8w+LmmHc+4VH+czIOfcMufcAufcggkTJgx+A4C8tWnFJnW0deioTx28q+aBZtbM1DvPvaOOfR1ZmJk/fA95A6zHiwtWBpPeeCV+Vh4tmwAAFI50KnmnSDrXzN6RdJ9ibZr/Jmm0mcV7iqZKqvd+rpd0iCR574+StKv3eB/3AChQtQ/UasTYETps8WGDXjuzZqY693XqvT++l4WZ+cPvNXkDHZ8QFwwl367JWXkAABSelEOec+5659xU59w0xTZOecY5938kPSvp095ll0p62Pv5Ee+1vPefcbHt8h6R9Flv983pkmZJejnVeQHIf13RLm34/QYdufRIFZcOvs/SYR86TMXB4iHVshkJR1RUWqTiYHr7SO3r2Kfd7bsTquSl0q5ZHYpV8thhEwCAwtH/Kv7UXSvpPjP7Z0mvSrrTG79T0q/MrE5Sk2LBUM65v5jZ7yS9KalT0tXOua4MzAtAnnj72bcVaY70eQB6XwLlAR22+DBtemJThmfmn0g4omBlsM+jIZKRyBl5ccHKoFoaWpJ6Pu2aAAAUHl9CnnPuOUnPeT9vVh+7Yzrn2iV9pp/7b5J0kx9zAZD/ah+oVaAioMM/fHjC98ysmakVX1uh5veaNerQURmcnT+i4aivZ+Rlql1z/MjxKrIidtgEAKCA+HFOHgAkrLurW+v/Z72O+PgRKilL/N+Zeo5SeGJotGzGK3np6qnkJdKuWZl8u2ZxUbEmlU+ikgcAQAEh5AHIqvf+9J7aGts0+/zZSd03/qjxqpxaqU3Lh0bLpl8hrz4cq+Ql1K7pVfJiy50TV1VRxZo8AAAKCCEPQFbVPlirkrISzTpnVlL3mZlm1MzQ5qc2q6sj/5ft+hbyWuo1snSkRgUHb1ENhAJy3U6d+zoHvba36lA1IQ8AgAJCyAOQNc45rX9wvWYsmaFARSDp+2fWzFQkHNGWVVsyMDt/+dmuOTk0OaENXOKfl+xZeVXlVbRrAgBQQAh5ALKmYXWDwlvCCe+qeaDDzzpcVmxD4igFPyt5iazHk2LtmvHPTkZ1qFrbW7erqzv/K6QAAGBwhDwAWVP7YK2KSop0xCeOSOn+stFlOuTkQ4bEujw/1+Qlsh5PirVrSkp685Wqiip1uS7t2rcr6fkBAID8Q8gDkBXOOdU+UKvpZ07XiDEjUn7OjJoZ2rp2q1q3t/b5/p72Pfr5mp/ntCrVFe1SZ3tn2iHPORdr16wY/PgEqVe7ZrKVPO+sPI5RAACgMBDyAGTFjjd2qKmuSUd9KrVWzbj4UQqbVvRdzfveH76nLz36JT20/qG0Picd8TVx6Ya8pn1NinRFEq7k9bRrJrsmr6JKEgeiAwBQKAh5ALKi9oFayaQjlx6Z1nOqj6/WyAkjtemJg0Ne074mLXtlmSTptpdvS+tz0hGvpKUb8uIHoSe8Js/7vGTbNatDXiWPHTYBACgIhDwAWVH7YK0OPfVQVUyqSOs5VmSauWSmNj2xSa57//Pgbl99u/Z27NWlx16qle+u1GvbXkvrs1LlW8jzzsibHEqsXTO+Ji/Zdk0qeQAAFBZCHoCM27Vxl3b8eUfarZpxM2pmqG1nm7au/aDytK9jn2596VadM/Mc/XjJjzWiZETOqnl+hbyGlgZJiR2ELqXerjmydKQqg5WsyQMAoEAQ8gBkXO2DtZKko87zKeR9ZIYk7XeUwn+u+081tjXq2lOu1dgRY3XxMRfrN3/+jXa1ZX/HSL/bNROt5JWWl0qWfLumFKvmbdtLJQ8AgEJAyAOQcesfXK/JCydr1KGjfHle+cRyVZ9Q3RPyOrs79aMXfqQTp5yoxYctliRds+gatXe2685X7/TlM5PhZ7vmhJETFChO7OB4M1MwFEy6XVOK7bBJJQ8AgMJAyAOQUc3vN6v+5fqUD0Dvz8yamdry4hbt271P9795v97e87auPeVamZkk6ehJR+tDh31It6++PevHKfjWrtnakHAVLy5YGUy6XVPyKnmsyQMAoCAQ8gBk1PqH1ktSRkKe63ba/NRm3fz8zTpy3JFaOnvpftf83Yl/p3eb39Xv3/q9r589mJ6QNyr9Sl6i6/HiAqGAouHk2zWrK6rZXRMAgAJByAOQUbUP1mrivIkad8Q4X5879aSpCo4K6oHnHtC6bev0jb/6hops/z/Szj3yXB1SeUjWN2CJhCOyIlPpyNK0nlPfUp/w8QlxwVDqlbzWaKtao30fMg8AAIYOQh6AjNm7Y6/e++N7mn3+bN+fXVRSpMM/fLh+3fFrVVdU6+JjLj7ompKiEv3twr/VM28/o7/s+Ivvc+hPJBxRsDLY0zqaimhXVDv27kipXTOVjVfiZ+XRsgkAwNBHyAOQMesfXi/X7TTnU3My8vyOMzu0ccpGXTn9SgVL+m6NvHL+lQoWB7NazYs0R9JejxcPW8lW8gKhQEobr3BWHgAAhYOQByBjah+o1ZgZYzTx6IkZef795fcr2B7UqZtP7fea8SPH63NHf06/ev1X2r1vd0bmcaB4JS8d8YPQk12Tl2q7ZnVFrJLHDpsAAAx9hDwAGdG+p11vP/22jjr/qLTaFvtT11SnR959RIvfWaztT2wf8NprFl2jto42/ee6//R9Hn3xJeR5Z+QlXcmrDKR8Tp5EJQ8AgEJAyAOQEW/971vq7uzWUZ/yd1fNuB+98COVFJXosimX6d2V7yq6t/9gc3z18TrlkFP0s9U/y8pxCn6EvIaWBkmJH4QeFz8nzzmX1H3jRo5TSVEJO2wCAFAACHkAMqL2gVqFpoQ0ZWFylahEbGvdprvX3a1Lj71UJy45UV3RLr3z3DsD3nPNomu0efdmPV73uO/zOZBf7ZqB4oDGjxyf1H3ByqC6O7vVFUkuzBZZkSaVT6KSBwBAASDkAfBddG9UdcvrYq2aRf63at760q2KdkX19b/6ug477TCVjChR3fK6Ae85/6jzNTk0OSsbsETCEQUqA2k9o76lXpNDk5NudQ2EAj1zSFZ1iLPyAAAoBIQ8AL6re7xOne2dvh+ALknhSFi3r75d5x91vo4Yd4RKyko0/Yzp2rR804D3lRaX6ksLvqQVm1Zow84Nvs9LkjraOrTunnVq29mmYCj9NXnJtmpK6vncVM/Ko5IHAMDQR8gD4LvaB2s1cvxIHXrqob4/e9kry9Qcada1p1zbMzajZoaa6prUVNc04L1XnXCVAsUB/fTln/o6p62vbtWjVz+qf538r3r4soc1ZvqYtNciNrQ0JL3piqSeNtGUzsqrqGZ3TQAACkBJricAoLB0Rjr11v++pbkXzlVRib//jhTtiuqWVbfojGlnaOGUhT3jM2tmSpLqnqjTopmL+r1/YvlEXTj3Qt392t266aybVBmsTHkukXBEf773z1r7H2u19ZWtKg4Wa+5n5ur4K4/XYYsPS2tHUeec6sP1+ujMjyZ9bzrtmlUVVWpsa1RXd5eKi4qTvh8AAOQHKnkAfLX5qc2KtkQz0qr5m9d/o4aWhv2qeJI0duZYjTl8zKAtm1JsA5bWaKvuXnd30p/vnNOWVVv08BUP61+r/1WPfvFRdUW7VHNrjb629Ws671fnadqHpqV9ZEQ4Etbejr2ptWtWpt6uWV1RrW7XrR17dyR9LwAAyB9U8gD4pvn9Zr34oxcVrAzq8LMO9/XZ3a5bP3zhhzp20rE6e8bZ+71nZppRM0Ov3fOaOiOdKgn2/0fbwikLdeKUE/XTl3+qLy/6sops8H/r2te0T6/96jW9+otXteONHSotL9W8z83TCZ8/QZMXJr85ymDixyckexC61GtNXoqVPCm2e2l1qDrp+wEAQH4g5AFIW8vWFv3x//5Ra5etlXNOH/nhR1Qc8Lfd7/cbfq/1O9frN+f/ps9QNbNmptbcvkbvP/++pp85fcBnXbPoGl380MVasWmFambW9HmNc07v/uFdrf2PtXrzgTfVFenS5IWT9fFlH9e8z85Le2OVgaR6ELr0QbtmSmvyvGC3tXWrjtfxSd8PAADyAyEPQMr27tirP938J625fY26O7t13N8cp8XfXqxRh47y9XOcc7r5+Zs1bfQ0XTD3gj6vmX7GdBWVFqlued2gIe8zcz+jr634mm57+baDQl5TXZPefOBNvXrnq2ra2KSy0WWa//n5mn/lfFUdW+XbdxpIfdgLealU8tJo1+xdyQMA5N6+3fu08vsrddo/nqaR40fmejoYQgh5AJLWtqtNL/zoBb1828vq3NepY/76GH3oux/SmMPHZOTz/vTen/Tilhd12zm3qaSo7z+2AhUBHXrqoapbXqeP/PAjAz4vUBzQF074gr6/8vvauHOjRr49Uuv/Z702PLxBjX9plCQdetqhWvydxZrz6TkqHVHq+3caSLxdM5U1eYHy9DZekcQOmwCQJ1b+80qtumWVujq69NHbkt+MC8MXIQ9Awtr3tOvFW17UqltWKdoa1dEXHa3F312s8UeOz+jn3vz8zRo/crwuP/7yAa+bWTNTT137lML1YVVO6X/nzK5ol85uPVs3uZt05Rev1JkPnCkrNh22+DDN//x8zV46W6Onjfb7aySsvqVeo8tGa2Rp8v9qa0WmQCiQUrtmWUmZRpeNppIHAHmg+b1mrf7ZapWOLNUr//6K/uprf5XT/zdhaCHkARhUpCWil259SS/+6EW172nXnE/P0Yf+6UOaOHdixj/7jR1v6NGNj+rG028cNPTEQ96mJzbp+Mv3X1PWvqddGx/fqA0Pb9DGxzYq2hLV3AvnavWc1frBJ36g4z5xnEaMHZHJr5Kw+pb6lNbjxQVDwZTaNSXvrLxWKnkAkGvP/dNzkqS/fuqvdc8Z9+i5f3pOn7z7k7mdFIaMlEOemR0i6ZeSJklykpY55/7NzMZK+q2kaZLekXSBc263xXZK+DdJH5XUJuky59xa71mXSvq29+h/ds7dk+q8APgnujeq1bev1vM3P699u/bpiE8coTO+d4aqjsvO2jRJ+uHzP9TI0pG6euHVg1478eiJqqiuUN3yOh1/+fFqfr9ZGx7ZoA0Pb9A7z76j7s5ulU8s19wL52r2J2frzNlnavGvF2vdnHU6eezJWfg2A2uNturmP92s5XXL+90QJhGBUEDRcPKVPCm2+QqVPADIrcY3G/XaPa/pxK+cqENOPkSLvrxIq25ZpVO+eYomzJmQ6+lhCEinktcp6WvOubVmFpL0ipk9KekySU87535gZtdJuk7StZLOkTTL+3WipDskneiFwhskLVAsLL5iZo8453anMTcAaehs79San6/Rn37wJ+3dvlcza2bq9BtP15RFqVeXUvFe83u69417dfXCqzVu5LhBrzczzayZqTf/+00tW7BMW1+JVaTGHTFOJ331JM3+5GxNPXGqrCi2O+csN0snVJ+g216+TV9c8EXfj0JIVFd3l+557R5965lvaVvrNl007yL9y0f+JeXnBStTr+RVVVRp1ZZVKX82ACB9z3zrGQUqAjrt+tMkSaded6peWfaKnv3Os7rggb43IAN6SznkOee2Strq/dxiZrWSpkhaKul077J7JD2nWMhbKumXzjknaZWZjTazau/aJ51zTZLkBcUaSfemOjcAqXv916/rqWufUktDi6afOV2nP3C6Dj3l0JzM5ccv/liS9NWTv5rwPXMvmKt1d69TcWmxzvrBWZq9dLbGz+57zaCZ6ZpF1+iyhy/T028/rQ8f/mFf5p2MZ95+Rl994qt6bftrOnnqyXrowod00tST0npmMBRMaeMVKdauua11m5xzOQu9ADCcvf/i+1r/P+t1xvfP6NlRc+T4kTr5ayfrD//0B9WvrteUhdn9R1cMPYOfApwAM5sm6XhJL0ma5AVASdqmWDunFAuA7/e6bYs31t94X59zlZmtMbM1jY2NfkwdQC/vv/C+HrrkIVUeUqlLnrlElzx9Sc4C3q62XfqPtf+hi+ZdpENHJT6HmTUz9e3It3XFi1fo1GtP7TfgxV0470KNHzlet718W7pTTsqGnRt07r3n6qxfnqU97Xt036fu0/OXP592wJOU8sYrUqyS19bRppZoS9rzAAAkxzmnp697WuWTynXSV/b//8HJXz1ZI8eP1DPfeiZHs8NQknbIM7MKSQ9I+opzLtz7Pa9q59L9jF7PW+acW+CcWzBhAv3IgJ862zv1yBWPaNQho3TJU5do+hkDnzWXabevvl1tHW365infTPre4tLED2IvKynTVfOv0u83/F5v73476c9K1q62Xfr7x/9e8+6Yp+feeU7/76z/p/VfXq8L513oW+UsWBlUe3N7SvdWV3gHonOMAgBkXd3yOr278l0t/s5iBSoC+70XDAV16vWnavOTm/X2s5n//xWGtrRCnpmVKhbwfuOce9Ab3u61Ycr77w5vvF7SIb1un+qN9TcOIItW3rRSO9fv1MeXffyg/7FkW1tHm259+VZ9bNbHNG/ivIx/3pcWfklFVqTbV9+esc+IdkX1k1U/0azbZumnq3+qK46/Qhuv2ajrTr1OZSVlvn7WxHkT1fxus/7wvT8kfS8HogNAbrhup6evf1pjDh+jEz5/Qp/XLPzbhaqcWqln/vEZxWopA9uxd4c+dPeH9N9/+W+/p4s8l3LI83bLvFNSrXPux73eekTSpd7Pl0p6uNf4/2/vvuPjqO6F/3/O7mq10qr3alvFTXLvNrGNCzbYYNMC3Dg4poTLTSAhT+5NBxK45OGXhCc3BAg39BBqSMCAbYwrAjds4yJL7rZ6bytpVbad3x8rrS3cJNlW8/fNa15Tdmb2jHyYne+ctlx5TQNsbdU61wILlFLhSqlwvJTgpAAAIABJREFUYEHbNiFEDynbV8aWJ7cw9jtjSV+Y3tvJ4ZU9r1DVVNWtUrzuSApJ4qaRN/HSnpdocjZd0nNrrfng0AdkPpfJj9b+iMmJk9l3/z6ev/55YoNiL3yCbpj+4+mM/c5YNj+6mc2/3typB4F28cFtJXkyjIIQQvSoA28foHxfOXMen4PRfPYaKSaLiVmPzKJoexFHPjpy3vO5PC7ueO8OsvKzuOfDe8iry7sMqRZ91cWU5F0F3AnMVUrtbZsWAU8C1yiljgLz29YBVgMngGPAC8D3ANo6XHkc2Nk2PdbeCYsQ4vLzuDx8eM+HBEQEsPD/Lezt5ODyuPjDtj8wLWkaMwfN7LHvfXDKg9S21PLG/jcu2Tm/Kv2KuX+by03v3ISfwY/V31rNJ8s+ueylkwajgSUvLWHcXeP47DefsfnRzgd6UpInhBA9z+1ws+nhTcSNi2PUHef/jRh/13gihkaw8Zcb0Z5z39sf3vgwm/I28ficxwFY8cEKPNpzSdMt+q6L6V3zC+BcDUjmnWV/DZx1oCut9cvAy91NixCi+7b9cRulu0v55j++eUkGA39t72t8dOSjs352tjZn6mu3kdqWWvLq8vjjwj/2aO+OMwfNZEzsGP785Z+5d8K9F/XdhbZCHtn8CK/tfY3IwEieXfQs9028D5PhYkat6RqD0cCSF5egDIqsx7PQHs2cx+dc8LrCLeGYjWZpkyeEED1o9wu7qT1Ry7I1y3zD/JyLwWRgzuNz+Ocd/yT7rWzGLBtzxj4rD63kyS1Pct+E+/jVrF+RGJzI3R/ezZ+2/4kfTf/R5boM0Yf03BOHEKLPqT5azeZHNjPiphGMvGXkRZ/vrey3WLFyBYNCBxFsDu7wmT5LH0znKl26acRNLBm+5KLT0xXtwyl896PvkpWfxewhs8+7v0d7KKov4mDlQQ5VHeJg1UEOVnmXK+wVmI1m/nPGf/KLmb8gzBLWQ1fRkTIobvjrDSiD4vMnPsfj9jDvt/POG+gppYgLiqPMLiV5QgjRExyNDrIey2Lw7MGkLUzr1DGZ38xky5Nb2PzIZjJvy+zQ4dixmmMs/2A5E+Mn8qfr/gTAinEr+ODwB/x8w89ZmL6QjOiMy3Itou+QIE+IK5T2aD767kcY/Y0senbRRZeafZ7/OStWrmDW4Fl8+u1P8Tf5X6KU9pxvjf4WP13/U57+8mlfkOdwOzhaffSMQO5Q1aEO7ffCLeGMjB7JDcNuYETUCG4eeTOp4am9dSk+yqC4/vnrUUbFlie3oD2a+U/OP++/d3xQvJTkCSFED9n2x23YK+zcsfKOTv8WK4Ni7hNzeXPxm+x5aQ+T7p8EeDsuu+XdWzAZTLx323u+jr2UUrxwwwuMem4Ud75/J9vu2YbZ2LudrInLS4I8Ia5Qu1/YTf5n+Sx5aQnB8cEXPuA8Dlcd5sZ3biQlLIX3b3+/XwZ4AIF+gdw7/l7+sO0PLHlrCYeqDnGi9gRu7fbtMyh0ECOjRjJzwkxGRo1kRNQIRkaPJDowus8OHq4MisXPLUYZFFt/txXt1lzz+2vOmd64oDhO1J7o4VQKIcSVp6mqia2/38qIm0aQNC2pS8emX5dO8lXJfPbYZ4z9zlhMFhP3f3w/2eXZrF62miFhQzrsH2ON4X+v/19ufvdm/jvrv3lszmOX8EpEXyNBnhBXoPqietb91zpS5qUw7q5xF3WuSnsli95chFEZWb1sNREBEZcolb3j+1O+z5sH3uR47XHGxI7h9szbfYHcsMhhBJmDejuJ3aKUYtEzi1AGxbantqE9mgVPLThroBcfFM+Wwi29kEohhLiyfP7bz3Hancx9Ym6Xj1VKMe//zuPVWa/y5TNfcmDOAV7f/zq/nv1rrk2/9qzH3DTyJr4z9jv89vPfsnjoYqYmTb3YSxB9lAR5QlxhtNZ8fP/HaLf2tte6iNKnZmczS95eQklDCZu/s7lPVE+8WINCB1H4o8LeTsZloZTiuqevw2A0sP2P29EezcI/LjwjD8QFxVHVVIXD7ZDqPEIIcZnYCmzsfHYnY1eMJXpkdLfOMXjmYNKvTeftl9/mr61/5dr0a3l49sPnPeZP1/6JTXmbWP7Bcvb8+x4C/QK79d2ib7uowdCFEP3PgbcPcHTVUeY+MZfw1PBun8ejPdz5/p3sKNrBGze/IW8D+wmlFAv/uJBpP5rGjj/tYM0P1pzRAU77EA8T/zqRlYdWdmmcPSGEEJ2z+dHNoODqX199UecZ++hYXr/2dcJd4fz9pr9jUOd/vA+1hPLq0lc5Un2En6776UV9t+i7JMgT4gpir7TzyQ8+IXFqIlMenHJR5/rpup/yz4P/5KkFT3HzyJsvUQpFT1BKseCpBUz/8XR2PrOT1Q+s7hDI3TzyZt659R0cbgc3vnMjM16ewaaTm3oxxUIIMbBU5FSw72/7mPLAFEKTQ7t9HrfHzY+P/Rh7iJ2bX78ZS6OlU8fNSZnDQ1Mf4pmdz7Du+Lpuf7/ouyTIu4SKvyzG45JBJkXftfahtbTYWljy4hIMxu7/7//czuf4w7Y/8MDkB3ho2kOXMIWipyiluOb31zDjJzPY9dwuVn1vlW9QXaUUt2XeRs73cnjhhhcoqi9i7t/msuD1Bewq2dXLKRdCiP5v4y83Yg4y842ff+OizvPYZ4/x6fFP+d3U3xGXH8fn//fzTh/723m/ZUTUCO5aeRe1zbUXlQ7R90iQd4k0ljXy6tWv8sLkFyjZXdLbyRHiDEdWHSH7zWxm/nImMaNiun2eVUdW8eCaB7lh2A38z7X/02d7lBQXppRi/pPzuepnV7H7+d18/B8f+wI9AJPBxL0T7uXog0d5asFTfFX6FZNfmMwt797CwcqDvZhyIYTovwq3FnJ45WFm/GQGgZHdbw+3+uhqHst6jBXjVvDQdQ8xbsU4dj23C1uBrVPHB/gF8PpNr1PWWMYPPvlBt9Mh+iYJ8i6RoLggbv77zTSWN/LilBdZ++O1OOyO3k6WEAC01rey6v5VxIyKYebPZ3b7PF+VfsXt793O+LjxvHXLWxgNxgsfJPo0pRTzfjuPmb+cyVd//YqP/v2jDoEegMVk4f9M/z+c+OEJfj3716w7vo5RfxnFXSvvIr8uv5dSLoQQ/Y/WmvU/W09QXBDTHprW7fOcrD3Jt//1bcbGjuW5Rc+hlGL2I97xXT977LNOn2dSwiQenvUwf9//d97Lfa/b6RF9jwR5l9DIm0fy/dzvM+G7E9j+/7bzl1F/4finx3s7WUKw/mfraShpYMlLSzCauxeYFdgKuP7N64kMjOSjf/sIq9l6iVMpeotSijmPz2HWw7PY8+IePrz3wzMCPYAQ/xAevfpRTvzwBA9NfYi3st9i2DPD+OGaH1Jhr+iFlAshRP9ybM0xCj4vYNYjszBbu9d7cYurhVv/cSse7eGft/2TAL8AAEIHhTLpe5PY+8peqg5Xdfp8v5j5CyYnTOb+j++ntKG0W2kSfY8EeZeYJczC9c9fz4qsFRj9jfx94d95/873sVfaeztp4gqVn5XPrr/sYupDU0mcktitc9habCx+czF2p53V31pNfHD8JU6l6G1KKeY8NofZj85m7yt7+deyf1G0vQi3033GvlGBUTy18CmOPniU5WOW8+zOZ0n9Uyq/2vgrbC2dqyYkhBBXGu3RbPj5BsLTwplw74Run+fB1Q/yVelXvH7T66RFpHX4bObPZ2IKMLH5kc2dPp+f0Y+/3fQ37E473/3ou9Kj8gCh+us/5KRJk/SuXX2rA4D1J9ZjVEZMBhNGgxHt1Bx49QD7XtmHf6A/s34yi4ylGZgMJt8+JoMJozL6loPMQRfs+laIznI2O3l+7PNot+b+/fd3662h0+1k0ZuL2Jy3mU+WfcK81HmXIaWiL8n67yw2PbIJNPhZ/Uiekczg2YMZMnsICZMTMPl3HGL1SPURHt70MO/mvEu4JZyffeNnPDDlARl7SQghTrP/jf28/+33ueWtWxh1x6hunePlPS9zz4f38Itv/IIn5j1x1n02PbKJrMezuG/3fcRP6PxL2ad3PM0PP/khL9zwAvdOuLdb6RM9Tym1W2s96YztEuRdGh7twfjYxbdPiguK44ZhN3DjiBuZmzIXi6lzXeEKcTbrf7aeLf/fFpZvWE7K3JQuH6+15t4P7+XlvS/zytJXWDFuxaVPpOiT7BV28rPyyfssj/zP8qnI9lbHNFlMJE1P8gV9iVMT8QvwA7xtNn+58Zd8cuwT4oLiWDp8KXOGzGFOyhxirN3v7EcIIfo7t8PNMyOewRJm4b5d96EMXe+0bE/pHqa/NJ1vDPoGa7+99pzt4ltsLTyd+jSJUxNZtnpZp8/v0R4WvL6AHcU72Hf/PlLDU7ucRtHzJMi7zLTWfFHwBW7txu1x4/K4Oiy73C4Of3qY7LezcWs3w24exuD5g/Hg8e3ncDvYWbKTNcfW0OhoxOpn5dr0a1k6fCmLhy0mIiCity9T9CMlu0t4ceqLjLtrHEteWNKtczyR9QS/2vQrHpn1CL+Z85tLnELRnzRVN1HweYEv6CvbWwYajGYjiVMTfUFf0vQktldt56ltT7Hp5CYaHA2Ad4D1OUPmMDdlLrMHzyY8ILyXr0gIIXrOl898yZoH17Dsk2WkL0zv8vG1zbVM/OtEnB4nX933FdHW6PPuv+X3W1j/k/WsyFrB4JmDO/09hbZCRv9lNKNjR7P5O5ulg7V+QIK8PqK+qJ5V31vFkY+OED8hnhteuOGMovRWVyub8jax8tBKPjzyISUNJRiVkZmDZ7J0+FKWDl9KSnjXS2XElcPtdPPC5BewV9j5fu73sYR1vUT4jf1v8O33v82dY+7ktRtfk6ESRActdS0UfHEq6Cv9qhTt1hhMBhImJzDk6iGkL0mnNLGUzfmb2XhyI18UfEGzqxmFYnz8eOYOmcuclDnMHDSTYP/g3r4kIUQf0VLXQsGWAvKz8inaVoRSCmusFWuM1TcPig3yrrdtMweZL+nvlPZonE1OHHYHBqOBgMiAbp/f0ejg6bSnic6MZvmG5V0+j0d7WPLWEj49/ilZd2UxLenCvXI6m538Of3PhKeGsyJrRZe+8/V9r7P8g+X8bv7v+K+r/qtLaRU9T4K8PkRrTe57uax5cA1NVU1M+9E05vxmDn6Bfmfs69EedpfsZuXhlXxw6ANyKnMAGB0z2hvwjVjKxPiJ8gAuOvj8t5+z8Zcbuf2D2xmxdESXj/8s7zMW/H0BM5JnsPbbazEbu9cDmLhytDa0Uril0Bf0lewswePyEDYkjIzbMhh1+yjCR4ezs2QnG09uZGPeRrYXbcfhdmBURqYkTvGV9M1InuHrLa63aK3RaGkjLUQPsFfYyf88n/ysfAqyCijb560pYPAzkDAxAaPZSGN5I/YKOy21LWc9h8li8gWApngTeUl55Ebkst+8H4dyEKNjiHHFEOWIIrIpkojGCMLqwwisDcRtd+NodOCwO3DanTgaHTibnB3Ob/Q3EpIUQmhyKCFJIYQkh3SYhyaHnjMQ/Oyxz9j86Gbu2X4PSVOTuvz3aa9V88x1z/D9Kd/v9HG7nt/Fqv9YxbdWfYuhi4Z2+jitNbf+41Y+PvIxu767i9Gxo7ucZtFzJMjrg5prm1n3k3XseXEP4anhXP+/15M6//z1n4/XHGfl4ZWsPLySLwq+wKM9JAYnsmT4Em/7l5Q58kB+has6VMXzY59nxI0juPWdW7t8/KGqQ8x4aQaxQbFsvXurVKsT3dJS18KhDw6R804OJ9afwOPyEJEe4Qv4YkbH0OxqZmvhVjad3MTGvI3sLN6JW7sxG81MT5rOvJR5XJN2DZMSJmEymC78pafRWnf55Zetxca6E+tYfXQ1a46todnZzI0jbuSbGd/kmrRr5N4qxCViK7SRn3UqqKs65O3u3xRgInl6MoNmDWLwrMEkTU064wW42+HGXmnHXmHHXu6d15XVsad6DztadrDHtIdjQcdwG9wYXUaSC5MJaA6gLqwOW6iNJmtTh/OZPCYiWiOIdkUTp+OIVbEkmBJI9E8kOTCZhKAElEvRUNxAfWE99UX12AptNBQ34HF5ANBotNIYA4wEJQX5JmuilcD4QDY9sYmYOTFc9YersLXasLXYzj4/x2et7laWjV7G6ze93qX7mtvp5tkRz+If4s99u7vWDrDSXsnov4wmLiiOHffuwN/k3+ljRc+SIK8Py/ssj4/v+5jqI9UMnjWYQTMHkTQ9ieTpyQREnPttdlVTFauOrGLl4ZWsPb6WJmcTVj8rs4fM5prUa5ifOp/M6Ewp5btCaK2pzKnkw3s/pOZoDd/L/R5BsUGdPj67PJtX9r7Ca/tew2Qwsf2e7VItWFwSTdVNHHrfG/Cd3HgS7dFEjYgi8/ZMMm/LJDrD27akvrWeLwq+8Jb0ndzInrI9gHd8vjlD5vjua8Mih3W4rzmbnFQcqKBsbxll+8oo31tO+f5yrLFWMr7pDSpjx8aecS/UWnOw6iCrj65m1dFVfFHwBS6PizBLGAvTFmIxWfjg0AfYWm2EWcJYOnwpt2XexvzU+RLwCdFJWmtqjtX4Arr8rHzq8uoA8A/xZ9A3BvmCuvZSu86cM7cyl/Un1rPh5AY2522mwdHgqwo+P2U+81LncVXSVahGhfZozEFmTBYTTc4m8m355Nflk1eX5122tS3X5VPa2HGcOKMyEmYJw6O9fSh4tAe3p+Pcg+ei/kZB5iBC/UMJtYR2nLctJwQncN/E+7rVY3F7j57hqeHev60CZVDe++EFlvfF7OOZic+w6OQi7nHeQ9KMJJJnJBM7OhaDqWMtB6fbSW1LLdVN1dQ011Dd7J3XNNdQ3VRNoF8g05KmMSVxilTPv8QkyOvjXC0utvx+C4dXHqZsbxna7f13iRoR5Q34ZiSTPCOZqBFRZ30T0+xsZsPJDaw5uob1J9dzpPoI4O2tc37qfN/DUUJwQo9el7i8tNaU7ysn971cct/LpfpwNcqguPnNmxl1+4W7Z65truWtA2/xyt5X2FWyCz+DHzcMv4Ffz/61VM8Ql4W9ws7Bfx0k550c8j7LAw0xo2K8Ad/tmUQOjfTtW9VUxcaTG1l/Yj3rTqwjry4PgHhzPONd40kvTCfuyzic2U7f4O3+If7Ejo0lZnQMtcdqObHhBNqtiRwWScZtGaTcnEJOYA5rjq1h9bHVvnOOiR3DovRFLBq6iOnJ030lhw63g/Un1vNuzrsdAr4bR9zIbRm3MS91ngR8QrTRHk3tiVrK9pZRuqeUsj1llH5Vir3cO1ZwYHQgg2cO9gV1sWNiMRg7VyW60FbIhpMbfIFdWWMZAOkR6cxLmcf81PnMGTKHyMDIC5zp/FpcLRTaCjsEgjXNNRgNRgzKgFG1zTuzjgFnoxMTJmLjYs8ayIX4h1zWzk20R7P5N5upPlwN2ruutT7nsva0rbctv5T2ElmxWSzYvYBmezPNAc20BrfijnPjDHPS5N9EPfXUO+rPmQajMuLW3jFXDcrAqJhRTE+a7p2SpzM0YmifK5DQWlPbUkteXR4na0+SV5fHNWnXMCZ2TG8n7QwS5PUjDruDkl0lFG4tpGhbEYVbC2mubga8g60nTUvyvk2Znkzi1ET8g88sQi+wFfgejNafWE9Vk7cqREZ0BvNT5nNN2jXMHjxb3qb0Q1prSneX+gK72uO1KKMiZU4KI28dyYgbR5y3BM/tcbPh5AZe3vMyHxz6gFZ3K2Nix3D3uLtZNmYZUYFRPXg14krWUNpA7nu55LyTQ+GWQgDixsWReXsmGbdm4Gp1Uba3jPJ95ZTuLSX3ZC45oTkcTzvOyZSTtAR42+ak6lRmRs9k8djFXDf5OoL8T+V/e6WdDf/YwHtb32M72zk55CQuPxcWbeHqxKu5acJNXJd+HcmhyRdMb6ur1Rvw5XoDvvrWesIt4d6AL/M25qXMw894ZttqIQYit8NNRU5bCfqetmlfGY4GBwDKqIjOiCZuXBzJVyUzZPYQIodH+h7mPdpDo6ORupY632RrsXVYr2upo7q5mm1F23wvr2OsMb6gbl7KPAaHdb7nSNF19a31TH1xKoeqDmFQBkIMIVgdVsw2M6ZqEwFNAQQ0BxBljSIxKZGU9BTSRqUxZOgQogKjiAiIIMQ/BFurjR1FO9hWtI1tRdvYXrSd+lZvYBgVGMW0pGm+wG9y4mSCzN77uMftobW+FUejo0O6OgSF6hzbT/vMZDEREH6qdpzWmrqWOm8QV+cN4r4+tfcO3e7P1/2ZB6Y8cJF/0UtPgrx+TGtNzdEaCrcV+gK/igMVoL3F6TGjYzqU9oWnhnfI5B7tYX/5ftYdX8f6k+vJys+ixdWCyWBiWtI0X9A3OWGyPKD0UdqjKf6y2BfY2fJtGEwGUualkPHNDEYsHUFg1PmrcRyrOcare1/ltX2vUVRfRLglnGWjl3HX+LsYHze+z71FE1eW+qJ6cv6RQ847ORTvKO7wmdFsJDrT+7AYNy6O2LGxRI2K4lDrId997YuCL3C4HfgZ/JiRPIP5qfOpa6lj1dFVHKo6BEBaSBqTnZNJ3JaI5WMLfi6/U6WIt2USOazzJQCtrlbWnVjHuznvsvLwSupb64kIiODG4d6Ab27KXLmfigGjtb7VG8y1B3R7y6jIqcDj9FZT9LP6ETc2juDxwehMjWOIg4bwBkqaSihuKKa2pfaM4K2+tR6PPn81R6uflTBLGOPixvkCu1Exo+T3qoc53U4aHY2EWkI7dEblsDso2ektlGh/Pm2u8RZKBEQEdHg2DR0cSqutlZa6FlrqWrDX2DlUc4ivGr4i25lNrsql2Oy99yutSKhJIKkwiYQTCSQXJhNeG45C4VEeXCZXh8ltdJ+x7evbHWYH7hFuHEMd1AbXUtBY4Asy2wWbg0kJT2FI2BCGhA7xztumlPAUwixhPfdH7wIJ8gaYFlsLxTuKKdxWSNHWIoq2F9Fa3wqANcZK8oxkX2lf/MR432DF4K2KsLVwq+/haHfJbjSaYHMwswbPYmL8RCbET2B8/HiSQ5LlZtpLtEdTuLWQ3PdyOfjPg9QX1WPwM5C2II2MWzMYvmT4edtsAjQ6Gnkv9z1e2fsKWflZGJSBBWkLuHvc3SwZvkQaUos+qS6vjiMfH8ESZvEGdCOiMPqdvzpTk7OJLwq+YP2J9aw/sZ49ZXswG83MHjybRUMXsXjoYoZGnupdrqG0gYP/9FYbLfiiAPCWImbclkHmbZlEpHV+XNJWVyufHv+Ud3PfZeWhlTQ4GogIiGDx0MVMiJ/AqJhRjI4ZTYw1Ru6nok9ztbioOlxFZU4lFTkV3vmBCmqP1+I0ObGF2nAMcngDucEOGqMaqQ2spdxVTmF94RkPzQZlID4onsjASMIsYYRZwgj1D/Utn209zBLmq8ooL0r6F+3RVB+p9gZ9bc+nlbmVFzzOP9QfS5gFd7Sb4uRi8mPyORZ2jKOWozQrb9Boxoy77b/u8nf6E1oTSrgtnMEhgxkzagyTZ0wmPS6dlDBvENcf79ES5A1wHreHytxKX/XOwq2F1BytAbxdEMdPiPe9TUmekUxwwqlqmtVN1WzK28S64+vIKsjicNVhNN58ERkQyfj48YyPa5vixzM0YqgMjnmZtNa3UrKrhIP/OsjBfx2ksbQRo7+R9GvTybg1g2HXD7vgmHdaa7YUbuGVPa/wbu67NDoaGRoxlLvG3cXysctJDEnsoasRovfUNNdgNpp9VX7Op764ntx/5JLzbg5F24oAiJ8YT9rCNMJTwgkdFEro4FBCk0PPOtTN6VpcLd6AL+ddPj3+KZVNpx5wogKjGB0z2hf0jY4dTWZ0Zp+sNu9xe6jIriAoLoiguM534NSu1dXK3rK9AExKmCS/GX2Mq9VF9eFqCrILOHbwGCdPnKSgpIDy+nIaAxtpCmzCHmTHGeOkKayJ2oBabNjOOE+MNYbkkGSSQ5MZFDKI5NBk33pySDLxwfFd7hlXDCzNNc0UbS+isbwRS5iFgPAALGEW32QONp+zXabb4yanModthds4VnMMs9GMxWTBYrLgb/L3zo3+ndoW6BdIqH8oFdkV7H9jPwfePEB9UT1+Vj9G3jSS0ctGkzo/9YwOZfoDCfKuQPZKe4egr2RnCa4WFwChg0N9AV/S9CTixsb5MrbdYWd/+X72lO1hT+kevir7igMVB3C4vfWhrX5WxsaNZXzceG+JX9x4MmMypfOBTtJa01jaSOXBSqoOVVF1sMo3byjx1v82WUwMXTSUkbeOZNj1w85od9noaKTQVkhhfSGFtkIKbAXe5fpCjlQfocBWQJA5iNsybuPu8XczI3lGv3w7JURPsxXYfNVGS3eX+jp0aRcYHegN+toDv0GhhA0O860HRgV2+H+twl5Bdnk2ByoOkF3hnR+oOIDdafftMyRsiC/wa58Pjxre4/fU+uJ6jn96nONrj3Ni/Qmaq5sxmAyMuHEEE++fSMrclHPeR8oay9hWuI2thVvZVrSNXSW7aHV7a5dEBUZxbfq1LB66mIVpCwf8sCwet4eqg1UUf1lMya4SnHYnBrMBo9mI0c/onZuNGPwMvmVMUGmqpIgiCnQBBe4C8l355DvyaXQ3YjG0Pbi2z40WAowB+Bv9vXOTPwGmACxGCxY/CwGmAAL8ArCYLDiaHRQWF1JSVUJFQwVVzipsBht2qx2Hv+Os12AxWIgNjiXGGtMhkDt9nhiSiMV0/peOQvRV2qPJ/zyf7Deyyf1HLi11LVhjrGTensnoZaNJnJLYb56bJMgTuB1uyvaW+YK+wi2FvqDCL9CPxCmJJE5NJGpEFJHDIokcFukb2NPhdnCw8mCHwG9v2V4aHY3e4w1+ZMZkMiZ2DJEBkQSZgzpMVj/rGdvapwC/gAE54LDH5aHmeM0ZgVzVoSpf1VoAc7CZ6JHRRI2IImpkFKEjQzFNNHmrv5wWyBXWnwrm6lrqOnyXQhEXFEdyaDKDQwcL6RavAAAQL0lEQVSzeOhibs24FavZ2tOXLcSA4Xa6aShpwJZvw1Zgoy6/DluBjfqCeu9yvu2MAZNNFpMv4AtJCiEwOhBrtJXAqEACowMJjAokICqASnMlh+2HyanMIbsim+zybA5XH8bl8b6IMxlMpIWnEegXiMlgOutkNBjP/pnyzv2MfsRaY72lLKGDGBQ6iKSQJF/w6Gx2kp+V7wvsKnO8pY5BcUGkLUgjZV4K5dnl7H15L801zUQOi2Tiv09k1PJRHHUe9QZ1RVvZVriNk3UnATAbzUyMn8j0pOnMSJ6Bw+1g1dFVrDm2xttDoTJy1aCrWDx0MYuHLiYjOqPfPEidjdaa+sJ6ir8s9k3tgR14e3u1hFtwO9x4nB4aDA2UB5VTGVpJZUQl1ZHVvsltOlUNzb/Fn6iqKCKrI7G0WHztipx+zg5tjk5fP3359HMBKI8iyB5EsCuYSFMkMdYY4iPjGZQwiMGDBhMXGucL6GKsMVj9rP3630WIrnC1uji25hj7/76fIx8fwd3qJiI9glHfGsWYZWO61F67N0iQJ87Q/uN0et3psr1lvsE9ASzhFm/ANzSSiGERvuAvcmgkJquJ4zXHfYHfnrI9HKg4gK3Vht1h91X57Iz2INBqtmIymFAoDMqAUuq8y0q1rZ+2HOofSnxQPPHB8SQEJ3RYjguKu6Rvx9tL5WqO1VB9tJqaYzXUHKmh8mAllScqsZvsNAd4uxwmEQwpBkgEd7QbR6iD5sBm6nU9NS2nxpL5em9OABEBEQwKHeR9i3ram9RBod7qMQnBCVKSKkQP01rTXNOMrcDWIRBsDwLri+ppqmrydU7xdUazsUPw5xflR21cLaXhpRQFFFFsKsZj9qDMCm3W3g4HPC7f5NbuDuu+7R7v9lZ361lfCEWoCMLqwwgoCiC4NphwezhpiWmMnTCWaXOnMXLiSAyGUy/eymvLeffdd1m7dS05KofixGKcZm8QEx8Uz4zkGb6gbkL8hLO29XV73Owo3sGqI6tYdXQV+8r3AfheSl0/7HquHnI1AX7nb2fc25prminZVdIhqGsfHsBoNhI3Lo6EKQkETAygOqWaE8YTHKo+xOHqwxyuOkx1c7XvXCaDibSwNNLD0kkPTSc9OJ1UayqpAalEGCLwuDy4W9143B60W3vnHn1q2e3t4v5sy06Xk1Z3Ky2uFsyBZlJHpRI9IhqTv1SdFOJ8WmwtHPznQbLfyObkppOgIWFyAqOXjWbUHaO6NP5wT5EgT3SK2+mmLq+O6iPVVB+ppuZojW+5vvBrvRAlBBM5LJKIoaeCv/C0cEwWEyhopZUmTxN2t51mTzN2tx27y06Tu4kmdxN2l903NTobsTvs2J12XB4XGo1He7zjtOAdu8WjPZ1armupo6ShhHJ7+Vl77ooMiOwYAAa1LQd7l61mqy8tdoedRkcjNVU1VJVWUV1ZTW11LXW2Ouob62lobqDV0IrTz4nD7MBpduIKdHkDO2PzOf/OBmUgIiCiwxQZEOmbJ4Uk+QK5pJAkKZETop/SWuNocGCvtNNU1URTZRNNVU2n1k/b1j5vqWs567n8Av18beSC4oKwxlk7rAfHB3u3x1h9A0pXl1ezfe12dm/bTe6hXMp1ObZQG00JTTRGN1JlqqJVt3b4HovJ4iv1K6ov8nVdb1RGRoWOYlDJIAI3BBJ/LJ5hQ4Yx+T8mM/pbo886nM+5FNoKfYPQbzi5gSZnEwGmAOalzvOV8nVmWIvLqcXWQmVOZYegrr2tO0DUyCjip8TjmeChekg1eZY89lftZ2/ZXkoaSnz7xQXFMTxyOMMihzE8cjjDo7zLKWEp0rGIEH1YfXE9B94+QPYb2ZTtKeO6P1/HlAem9HayziBBnrhoziant8TqSLW31OrIqQCwqarp4r9AgcFowGAyYPDzzo1+xvMuG/2Mvv3bl9vbOmAGu8WOzWKj1q+WOr86ak211BhrqDHUUEMN1VRT5anqUm9NRrcRs9uMBQuBxkCC/IMIDggmOCiY0JBQQiwhRFjaArfAyLMGc8H+wQOyiqoQ4uK5nW6aq5uxV9hpLG+ksezUZC+zd1hv76786wIivZ0b1J6oBe0dYzV1fippC9NIW5BG6KBQwBuEVjdXU2ArOOsUY41hRvIMZiTPYFLCJAL9vEO1tDa0kv1mNrv+sovyfeWYg82M+fYYJt0/idgxsV263hZXC5vzNrPqyCo+Pvqxb4D60TGjmZcyjyFhQzq0B4uxxlzS+2dzbTOVuZXeKafSt9xQfKpWRXBiMNFTo2md2ErVkCryAvM4UHOAfeX7fM0WjMpIRnQG4+LG+aaxsWMvenBuIUTvq8ytJCg+qMNYe32FBHnismquaab6aDV1J+twO91nr1rSmWW3xuPyeKupOL1tGDwuDx5nx/ULLbsd7lOT091h/eu1SD3KQ3NAMw3BDTQGNeL0c+Lv8ScyOpLohGhikmOIHxJPQloCCcMTiBoc1S97XxJCDDyuVpc3GCztGAw2ljXSVNlE9Kho0hemkzA54Zw92F0MrTXFO4rZ9fwuct7JwdXiInlGMhPvn0jmNzO9NTu6eL6DVQd91Tp3FO+gxdWxZNPP4EdiSGKHqutJIUkd1qMCo85oU9ZU1eQL4EpzSik+XEzJ8RJqbbW+2hg6SOOf6o//IH9MCSaIgeLAYnLqczhYeRC39r4QDDYHMzZuLONiTwV0mTGZ0hGJEKLHSZAnRBuP+1QQeLaA0C/Qj9DkUAnkhBCiC5prmtn72l52P7+b6iPVBEQEkH5dOkHx3iqkQbHeuTXGijXWijX6VLXSc2kvafx6J1RF9UW+9aL6Ipyejh3g+ONPFFGEOEJobmmmydVEi7HFF8x9vWOS80kMTvQFcuPjxjMubhwp4SlSG0MI0SdIkCeEEEKIy05rTd6mPHY9v8vXMUn78D1fZwm3dAgAA2MCTwWDsVb8Av3ObMtY2XHdXmPHbrVTH1JPfUg9tlCbb7k5rJlASyAh1hBCQ0OJiIggKiaK8MjwDj0/W81WrH5WrGarb3v7cmfGWhRCiN5yriBPulkSQgghxCWjlCJlbgopc1OAts5nGh3Yy+3YK+y+tob2CnuHbRUHKrBX2M/ZzlAZFYGRgb5hKWJGx/h6J7VGW886XIXRTwZhF0JcmSTIE0IIIcRlo5TCP9gf/2B/ItIjLri/2+GmqaqJxvJGnE1OAqO8wZslzIIyyNhtQgjRGX0myFNKXQv8CTACL2qtn+zlJAkhhBCihxnNRoITgglOCO7tpAghRL/VJ1oNK6WMwLPAdUAG8G9KqYzeTZUQQgghhBBC9D99IsgDpgDHtNYntNYO4G1gaS+nSQghhBBCCCH6nb4S5CUChaetF7Vt60ApdZ9SapdSaldlZWWPJU4IIYQQQggh+ou+EuR1itb6r1rrSVrrSdHR0b2dHCGEEEIIIYToc/pKkFcMJJ+2ntS2TQghhBBCCCFEF/SVIG8nMFQplaKUMgN3AB/2cpqEEEIIIYQQot/pE0MoaK1dSqkHgLV4h1B4WWud08vJEkIIIYQQQoh+p08EeQBa69XA6t5OhxBCCCGEEEL0Z32luqYQQgghhBBCiEtAgjwhhBBCCCGEGEAkyBNCCCGEEEKIAUSCPCGEEEIIIYQYQCTIE0IIIYQQQogBRII8IYQQQgghhBhAJMgTQgghhBBCiAFEaa17Ow3dopSqBPJ7Ox3irKKAqt5OhLiiSR4UfYHkQ9EXSD4UfYHkw8tnsNY6+usb+22QJ/oupdQurfWk3k6HuHJJHhR9geRD0RdIPhR9geTDnifVNYUQQgghhBBiAJEgTwghhBBCCCEGEAnyxOXw195OgLjiSR4UfYHkQ9EXSD4UfYHkwx4mbfKEEEIIIYQQYgCRkjwhhBBCCCGEGEAkyBNCCCGEEEKIAUSCPHFBSqmXlVIVSqkDp237vVLqkFJqv1LqfaVUWNt2P6XUa0qpbKXUQaXUz0875lql1GGl1DGl1M9641pE/9XFfGhWSr3Slg/3KaWuPu2YiW3bjymlnlZKqV64HNFPnSMfPt6WB/cqpT5VSiW0bVdteexY2+cTTjvmO0qpo23Td3rjWkT/1cV8OEIptU0p1aqU+s+vnUd+l0W3dTEfLmvbnq2U2qqUGnvaMZIPLwMJ8kRnvApc+7Vt64BRWusxwBGgPZj7JuCvtR4NTAT+XSk1RCllBJ4FrgMygH9TSmX0ROLFgPEqnc+H3wVoy4fXAE8ppdrvd39p+3xo2/T1cwpxPq9yZp75vdZ6jNZ6HPAx8Ejb9us4lc/uw5v3UEpFAI8CU4EpwKNKqfDLn3QxgLxK5/NhDfAD4A+n7yy/y+ISeJXO58OTwOy23+XHaeuIRfLh5SNBnrggrXUW3h+J07d9qrV2ta1uB5LaPwKsSikTEAA4gHq8DzLHtNYntNYO4G1gaU+kXwwMXcyHGcDGtn0qgDpgklIqHgjRWm/X3l6n/gbc2BPpFwPDOfJh/WmrVrz3QfDe4/6mvbYDYW15cCGwTmtdo7WuxfuyQl42iE7rSj7UWldorXcCzq+dRn6XxUXpYj7c2na/g46/15IPLxNTbydADAh3A++0Lb+H93/OUiAQ+JHWukYplQgUnnZMEd632EJcKqfnw33AEqXUW0Ay3lLlZMCDN++1KwISezKRYmBSSj0BLAdswJy2zWe77yWeZ7sQF+Uc+fBc5HdZXBadyIf3AGvaliUfXiZSkicuilLql4ALeKNt0xTADSQAKcCPlVKpvZQ8cYU4Sz58Ge8PxS7gf4CtePOlEJeF1vqXWutkvHnwgd5Oj7gyST4UfcH58qFSag7eIO+nvZG2K4kEeaLblFIrgOuBZfrUgIvfAj7RWjvbqsltASYBxXhLUtoltW0T4qKcLR9qrV1a6x9prcdprZcCYXjb7BVzqooISD4Ul94bwC1ty+e678n9UFxup+fDc5F8KC63DvlQKTUGeBFYqrWubtss+fAykSBPdItS6lrgJ8ASrXXTaR8VAHPb9rEC04BDwE5gqFIqRSllBu4APuzZVIuB5lz5UCkV2Jb/UEpdA7i01rla61KgXik1ra1XzeXAyt5Iuxg4lFJDT1tdiveeB9573PK2XjanAba2PLgWWKCUCm/rcGVB2zYhuu08+fBc5HdZXHLnyodKqUHAv4A7tdZHTttH8uFlIm3yxAW1tWu6GohSShXh7RXu54A/sK6tB/rtWuv78faQ9IpSKgdQwCta6/1t53kA74OMEXhZa53T09ci+q8u5sMYYK1SyoP3jeCdp53qe3h7BAvA2yZgDUJ00jny4SKl1HC8bT7zgfvbdl8NLAKOAU3AXQBt7ZQfx/twA/CY1rpD5wVCnE9X8qFSKg5v1fUQwKOUegjI0FrXy++yuBhdvB8+AkQCz7X9Xru01pO01i7Jh5eHOlXLTgghhBBCCCFEfyfVNYUQQgghhBBiAJEgTwghhBBCCCEGEAnyhBBCCCGEEGIAkSBPCCGEEEIIIQYQCfKEEEIIIYQQYgCRIE8IIYQQQgghBhAJ8oQQQgghhBBiAPn/AZ0BpN9SxF4aAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 1080x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LtU9j1Rsij49",
        "outputId": "f28265ee-c280-4f5c-8352-2e764f6d010f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        }
      },
      "source": [
        "predicted_df=pd.DataFrame()\n",
        "predicted_df['load_date']=AnomaliesData['timestamp'][-48:]\n",
        "predicted_df['actuals']=test\n",
        "predicted_df['predicted']=predictions\n",
        "predicted_df.reset_index(inplace=True)\n",
        "del predicted_df['index']\n",
        "predicted_df.head()"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>load_date</th>\n",
              "      <th>actuals</th>\n",
              "      <th>predicted</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2015-04-21 03:00:00</td>\n",
              "      <td>1334</td>\n",
              "      <td>1102.542009</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2015-04-21 04:00:00</td>\n",
              "      <td>847</td>\n",
              "      <td>1092.374260</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2015-04-21 05:00:00</td>\n",
              "      <td>675</td>\n",
              "      <td>856.415241</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>2015-04-21 06:00:00</td>\n",
              "      <td>502</td>\n",
              "      <td>651.847114</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>2015-04-21 07:00:00</td>\n",
              "      <td>389</td>\n",
              "      <td>504.849428</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "            load_date  actuals    predicted\n",
              "0 2015-04-21 03:00:00     1334  1102.542009\n",
              "1 2015-04-21 04:00:00      847  1092.374260\n",
              "2 2015-04-21 05:00:00      675   856.415241\n",
              "3 2015-04-21 06:00:00      502   651.847114\n",
              "4 2015-04-21 07:00:00      389   504.849428"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WaKgG3pGij5A"
      },
      "source": [
        "What we have now is the real and predicted data, we just need to look at the average and distribution of the 'error' of our model to create potential approximate 'normal' pattern.\n",
        "\n",
        "Subsequently, anything outside of that pattern is a potential anomalies."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-yfnWCIWij5A"
      },
      "source": [
        "import numpy as np\n",
        "def anomalies(df,rollingwindows):\n",
        "    df['error']=df['actuals']-df['predicted']\n",
        "    df['meanerror'] = df['error'].rolling(window=rollingwindows).mean()\n",
        "    df['deviationerror'] = df['error'].rolling(window=rollingwindows).std()\n",
        "    df['-3sderror'] = df['meanerror'] - (3 * df['deviationerror'])\n",
        "    df['3sderror'] = df['meanerror'] + (3 * df['deviationerror'])\n",
        "    df['-2sderror'] = df['meanerror'] - (2 * df['deviationerror'])\n",
        "    df['2sderror'] = df['meanerror'] + (2 * df['deviationerror'])\n",
        "    df['-1sderror'] = df['meanerror'] - (1 * df['deviationerror'])\n",
        "    df['1sderror'] = df['meanerror'] + (1 * df['deviationerror'])\n",
        "    cut_list = df[['error', '-3sderror', '-2sderror', '-1sderror', 'meanerror', '1sderror', '2sderror', '3sderror']]\n",
        "    cut_sort = np.sort(cut_list.values)\n",
        "    df['Anomalies_Strength'] = [(lambda number: np.where(cut_sort == df['error'][number])[1][0])(number)\\\n",
        "                                for number in range(len(df['error']))]\n",
        "    severity = {0: 'Severe Anomaly', 1: 'Anomaly', 2: 'Unusual', 3: 'Normal', 4: 'Normal',\n",
        "                5: 'Unusual', 6: 'Anomaly', 7: 'Severe Anomaly'}\n",
        "    df['Anomalies_Strength'] =  df['Anomalies_Strength'].map(severity)\n",
        "    df['anomaly_points'] = np.where(df['Anomalies_Strength'] == 'Severe Anomaly', df['error'], np.nan)\n",
        "    df = df.sort_values(by='load_date', ascending=False)\n",
        "    df.load_date = pd.to_datetime(df['load_date'].astype(str), format=\"%Y-%m-%d\")\n",
        "    df[-rollingwindows:]['anomaly_points'] = np.NaN\n",
        "    df[-rollingwindows:]['Anomalies_Strength'] = 'Normal'\n",
        "    df['upper_boundary'] = df['predicted']+df['1sderror']\n",
        "    df['lower_boundary'] = df['predicted']+df['-1sderror']\n",
        "    df['upper_boundary_high'] = df['predicted']+df['3sderror']\n",
        "    df['lower_boundary_high'] = df['predicted']+df['-3sderror']\n",
        "    df = df.drop(columns=['1sderror', '-1sderror','2sderror','-2sderror','-3sderror','3sderror'])\n",
        "    return df"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QcjJxTmzij5C"
      },
      "source": [
        "classify_df=anomalies(predicted_df,12)"
      ],
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3gTNKZ1Yij5E",
        "outputId": "faa77302-1f45-4cfa-c304-e8997ab385af",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 474
        }
      },
      "source": [
        "ggplot(classify_df)+ geom_line(aes(x='load_date',y='actuals')) \\\n",
        "+ geom_point(aes(x='load_date',y='actuals',color='Anomalies_Strength'))+\\\n",
        "  scale_color_manual(values = {'Normal':\"black\",'Anomaly':\"orange\",'Severe Anomaly':\"red\",'Unusual':'yellow'})\\\n",
        "+theme(axis_text_x = element_text(size = 5.5,angle = 90, hjust = 20, vjust = -400))\\\n",
        "+geom_line(aes(x='load_date',y='upper_boundary'),color='orange',linetype=\"dotted\")\\\n",
        "+geom_line(aes(x='load_date',y='lower_boundary'),color='orange',linetype=\"dotted\")\\\n",
        "+geom_line(aes(x='load_date',y='upper_boundary_high'),color='red',linetype=\"dotted\")\\\n",
        "+geom_line(aes(x='load_date',y='lower_boundary_high'),color='red',linetype=\"dotted\")"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<ggplot: (-9223363256731597431)>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "frgudNZcij5G",
        "outputId": "2391ae52-863c-4a46-9227-ac7f566717b2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 289
        }
      },
      "source": [
        "classify_df.head()"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>load_date</th>\n",
              "      <th>actuals</th>\n",
              "      <th>predicted</th>\n",
              "      <th>error</th>\n",
              "      <th>meanerror</th>\n",
              "      <th>deviationerror</th>\n",
              "      <th>Anomalies_Strength</th>\n",
              "      <th>anomaly_points</th>\n",
              "      <th>upper_boundary</th>\n",
              "      <th>lower_boundary</th>\n",
              "      <th>upper_boundary_high</th>\n",
              "      <th>lower_boundary_high</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>47</th>\n",
              "      <td>2015-04-23 02:00:00</td>\n",
              "      <td>445</td>\n",
              "      <td>781.473252</td>\n",
              "      <td>-336.473252</td>\n",
              "      <td>37.698309</td>\n",
              "      <td>543.889056</td>\n",
              "      <td>Normal</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1363.060618</td>\n",
              "      <td>275.282505</td>\n",
              "      <td>2450.838730</td>\n",
              "      <td>-812.495608</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>46</th>\n",
              "      <td>2015-04-23 01:00:00</td>\n",
              "      <td>717</td>\n",
              "      <td>685.537941</td>\n",
              "      <td>31.462059</td>\n",
              "      <td>69.818659</td>\n",
              "      <td>531.011939</td>\n",
              "      <td>Normal</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1286.368539</td>\n",
              "      <td>224.344661</td>\n",
              "      <td>2348.392417</td>\n",
              "      <td>-837.679218</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>45</th>\n",
              "      <td>2015-04-23 00:00:00</td>\n",
              "      <td>718</td>\n",
              "      <td>662.786431</td>\n",
              "      <td>55.213569</td>\n",
              "      <td>66.042498</td>\n",
              "      <td>531.470421</td>\n",
              "      <td>Normal</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1260.299349</td>\n",
              "      <td>197.358508</td>\n",
              "      <td>2323.240191</td>\n",
              "      <td>-865.582334</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>44</th>\n",
              "      <td>2015-04-22 23:00:00</td>\n",
              "      <td>670</td>\n",
              "      <td>833.300994</td>\n",
              "      <td>-163.300994</td>\n",
              "      <td>58.816624</td>\n",
              "      <td>532.219965</td>\n",
              "      <td>Normal</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1424.337583</td>\n",
              "      <td>359.897653</td>\n",
              "      <td>2488.777513</td>\n",
              "      <td>-704.542277</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>43</th>\n",
              "      <td>2015-04-22 22:00:00</td>\n",
              "      <td>843</td>\n",
              "      <td>778.259680</td>\n",
              "      <td>64.740320</td>\n",
              "      <td>80.837811</td>\n",
              "      <td>527.641327</td>\n",
              "      <td>Normal</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1386.738818</td>\n",
              "      <td>331.456164</td>\n",
              "      <td>2442.021472</td>\n",
              "      <td>-723.826490</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "             load_date  actuals  ...  upper_boundary_high  lower_boundary_high\n",
              "47 2015-04-23 02:00:00      445  ...          2450.838730          -812.495608\n",
              "46 2015-04-23 01:00:00      717  ...          2348.392417          -837.679218\n",
              "45 2015-04-23 00:00:00      718  ...          2323.240191          -865.582334\n",
              "44 2015-04-22 23:00:00      670  ...          2488.777513          -704.542277\n",
              "43 2015-04-22 22:00:00      843  ...          2442.021472          -723.826490\n",
              "\n",
              "[5 rows x 12 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3Fqmulebij5I"
      },
      "source": [
        "# Looking into the past - Kleinberg's Burst Detection"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4McCvnlGij5I"
      },
      "source": [
        "http://www.cs.cornell.edu/home/kleinber/bhs.pdf\n",
        "\n",
        "Kleinberg’s algorithm identifies periods of time in which an event of particular interest is uncharacteristically frequent, or “bursty.”  It can be used to detect bursts in a continuous stream or batches of events by assumeing that there are only two possible states, the baseline state and the 'bursty' state. \n",
        "\n",
        "The probability of event of interest happening in baseline state is equal to the total number of event of interest divided by total number of events. The 'bursty' state probability is constant multiplied by baseline probability."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mtK7JrQoij5J"
      },
      "source": [
        "\n",
        "* pybursts (Continuous stream) = https://github.com/romain-fontugne/pybursts (Forked version which fix some problem)\n",
        "* burst_detection (Batch event) = https://github.com/nmarinsek/burst_detection"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nzMudvojij5J"
      },
      "source": [
        "The total cost of transitioning from normal state to burst is equal to the sum of the two functions below. With the cost found, we can then find the optimal state sequence that tell us what is the state at each timepoint by measuring if the new state has been reached or not."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MD9IVmatij5K"
      },
      "source": [
        "The goodness of fit between the observed proportion and the expected probability of each state.\n",
        "\n",
        "$\\sigma= - ln[\\binom{d_t}{r_t}p_i^{rt}(1-p_i)^{d_t-r_t}]$"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IuHlF_Qo7gUx"
      },
      "source": [
        "import sympy\n",
        "\n",
        "def fitcost(normal,interest,p):\n",
        "    return -np.log(np.float(sympy.binomial(normal,interest))*(p**interest)*((1-p)**(normal-interest)))"
      ],
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pdB4BHnZij5K"
      },
      "source": [
        "The difficulty of transitioning from the previous state to the next state.\n",
        "\n",
        "$\\tau= - (i_{next}-i_{prev})*\\gamma*ln(n)$"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O2rDwCt97GkD"
      },
      "source": [
        "def transitioncost(previousstate,nextstate,gamma,n):\n",
        "    if previousstate>=nextstate:\n",
        "        return 0\n",
        "    else: \n",
        "        return (nextstate-previousstate) * gamma * np.log(n)"
      ],
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KzdP7GpG9MKk"
      },
      "source": [
        "def burstdetection(interest,normal,totaltimepoint,activestatecost,gamma,smooth):\n",
        "    \n",
        "    k = 2 \n",
        "    # smooth time_series, can be ignored if not needed.\n",
        "    if smooth > 1:\n",
        "        temp_p = interest/normal\n",
        "        temp_p = temp_p.rolling(window=smooth, center=True).mean()\n",
        "        interest = temp_p*normal\n",
        "        real_n = sum(~np.isnan(interest)) \n",
        "    else: \n",
        "        real_n = totaltimepoint\n",
        "          \n",
        "    #calculate the expected proportions for states 0 and 1\n",
        "    p = {}\n",
        "    p[0] = np.nansum(interest) / float(np.nansum(normal)) #baseline state\n",
        "    p[1] = p[0] * activestatecost  #cost to move from normal state to a burst state. \n",
        "    if p[1] > 1: #active state can't be more than maximum of proportuib\n",
        "      p[1] = 0.99999\n",
        "\n",
        "    #initialize matrices to hold the costs and optimal state sequence\n",
        "    cost = np.zeros([totaltimepoint,k])\n",
        "    q = np.zeros([totaltimepoint,1])\n",
        "\n",
        "    for t in range(int((smooth-1)/2),(int((smooth-1)/2))+real_n):\n",
        "\n",
        "        for j in range(k): # There are two states, which one has less cost: 'Active' or Normal\n",
        "            if t == int((smooth-1)/2):\n",
        "                cost[t,j] = fitcost(normal[t],interest[t],p[j])\n",
        "            else:\n",
        "                cost[t,j] = transitioncost(q[t-1],j,gamma,real_n) + fitcost(normal[t],interest[t],p[j])\n",
        "\n",
        "        #add the state with the minimum cost to the optimal state sequence\n",
        "        q[t] = np.where(cost[t,:] == min(cost[t,:]))\n",
        "\n",
        "    return q,p"
      ],
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9zrbIAzCij5K"
      },
      "source": [
        "AnomaliesData_Start = df[df['key']=='AAPL']\n",
        "min_max_scaler = preprocessing.MinMaxScaler()\n",
        "np_scaled = min_max_scaler.fit_transform(AnomaliesData_Start[['value','total']])\n",
        "AnomaliesData = pd.DataFrame(np_scaled*100, columns = ['value','total']) \n",
        "#Scaled the data because of issue with binomial coefficient\n",
        "r = pd.Series(AnomaliesData['value']).astype('float')\n",
        "d = pd.Series(AnomaliesData['total']).astype('float')"
      ],
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dR6fSCDwij5M"
      },
      "source": [
        "Burst Detection Function: burst_detection(r,d,len(r),s,d,smooth_win)\n",
        "\n",
        "* r = number of events of interested\n",
        "* d = number of all events\n",
        "* s = distance between normal state and a burst state\n",
        "* gamma = cost to move from normal state to a burst state.\n",
        "* smooth = smoothing windows\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EWZQ3U0Oij5N"
      },
      "source": [
        "[q,p] = burstdetection(r,d,len(r),2,1,smooth=3)\n",
        "bursts = bd.enumerate_bursts(q, 's='+str(3)+', g='+str(1))\n",
        "bursts = bd.burst_weights(bursts, r, d, p)"
      ],
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AUS4Pa1Fij5O"
      },
      "source": [
        "data_r = pd.DataFrame({'day':range(len(r)),'data':r,'total':d,'outlier':AnomaliesData_Start['outlier']})\n",
        "data_r['Prediction'] = 0\n",
        "for index, row in bursts.iterrows():\n",
        "    data_r['Prediction'] = np.where(data_r['day']>=row.begin,np.where(data_r['day']<=row.end,1,data_r['Prediction']), data_r['Prediction'])"
      ],
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_PRy1cUnij5R",
        "outputId": "21dc24ae-79a8-4463-bc00-2b719c745d28",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 461
        }
      },
      "source": [
        "ggplot(data_r, aes(x='day',y='data'))+ geom_line()+ theme(\n",
        "  axis_text_x = element_text(size = 5.5,angle = 90, hjust = 20, vjust = -400)\n",
        ")+geom_point(data_r[(data_r['outlier']==1)],aes(x='day',y='data'), color='red',shape = 7)+geom_point(data_r[(data_r['Prediction']==1)],aes(x='day',y='data'), color='blue',shape = 6)"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<ggplot: (8780122508015)>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DkLhV6epij5S"
      },
      "source": [
        "The strength, or weight, of a burst (that begins at time point t1 and ends at time point t2) can be estimated with the following function:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8nXbGldNij5T"
      },
      "source": [
        "def plot_burst_timeline(bursts, timepoints):\n",
        "    \n",
        "    f, ax = plt.subplots(figsize=(8, 0.5))\n",
        "    ax.set(xlim=(0, timepoints), ylabel=\"\", xlabel=\"\")\n",
        "\n",
        "    #create boxes around bursting periods\n",
        "    for index, burst in bursts.iterrows():\n",
        "\n",
        "        #define outline positions\n",
        "        y = 0.25\n",
        "        xstart = burst['begin'] -1 \n",
        "        width = burst['end'] - burst['begin'] +1\n",
        "\n",
        "        ax.add_patch(Rectangle((xstart,y), width, height=0.5, \n",
        "                               facecolor='#00bbcc', edgecolor='none', linewidth=1))\n",
        "\n",
        "    ax.xaxis.set_visible(False)\n",
        "\n",
        "    plt.axhline(0.5, linewidth=1, color='k', alpha=1, zorder=0.5)\n",
        "    plt.show()"
      ],
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1yWHg5w1ij5V",
        "outputId": "12f744f0-09ca-433f-f583-52863e785d4d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "plot_burst_timeline(bursts, len(r))"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x36 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0ncyQmT8ij5X"
      },
      "source": [
        "# Finding the Pattern - CNN Autoencoders "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ERS_3vEHij5X"
      },
      "source": [
        "Last, but by no mean least. There might be some situation where you lack understanding of the data to be able to correctly estimated boundary or level of energy for either forecasting or burst detection.\n",
        "\n",
        "Essentially, Autoencoder try to reconstruct the data based on what is learn before. If the autoencoder is good enough, it should be able to generalized to most situation and retain most of the standard information of time-series."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BfSITs2Uij5X"
      },
      "source": [
        "import time\n",
        "import numpy as np\n",
        "import torch\n",
        "import torch.nn.functional as F\n",
        "from torchvision import datasets\n",
        "from torchvision import transforms\n",
        "from torch.utils.data import DataLoader\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "if torch.cuda.is_available():\n",
        "    torch.backends.cudnn.deterministic = True\n",
        "    \n",
        "def get_value_from_df(df):\n",
        "    return df.value.to_list()\n",
        "\n",
        "def normalize(values,mean=None,std=None):\n",
        "    if mean is None:\n",
        "        mean = np.mean(values)\n",
        "    values -= mean\n",
        "    if std is None:\n",
        "        std = np.std(values)\n",
        "    values /= std\n",
        "    return values, mean, std\n",
        "\n",
        "def create_sequences(values, time_steps=24):\n",
        "    output = []\n",
        "    for i in range(len(values) - time_steps):\n",
        "        output.append(values[i : (i + time_steps)])\n",
        "    # Convert 2D sequences into 3D as we will be feeding this into\n",
        "    # a convolutional layer.\n",
        "    return np.expand_dims(output, axis=2)"
      ],
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NFMMYLQYij5Z"
      },
      "source": [
        "Using the same data and splitting into test and train, we then decided to use 24 hours of data for each autoencoders. Hence, 24 timesteps."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j-OFBABjij5a"
      },
      "source": [
        "Autoencoder_Data = df[df['key']=='AAPL']"
      ],
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4eBbvpxXij5d"
      },
      "source": [
        "# Get the `value` column from the training dataframe.\n",
        "training_value = get_value_from_df(Autoencoder_Data)\n",
        "# Normalize `value` and save the mean and std we get\n",
        "training_value, training_mean, training_std = normalize(training_value)\n",
        "TIME_STEPS = 24\n",
        "batch_size = 128\n",
        "train = create_sequences(training_value,24)"
      ],
      "execution_count": 39,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jlQVVnOTij5f"
      },
      "source": [
        "train, test = (train[0:1000,:,:], train[1000:,:,:])"
      ],
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wqgbAEFmij5h",
        "outputId": "8bbc48f8-fede-4719-9650-32e84860a37b",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "X_train = torch.FloatTensor([train]).cuda().reshape(-1, 24, 1)\n",
        "X_test = torch.FloatTensor([test]).cuda().reshape(-1, 24, 1)\n",
        "print(\"training input shape: \", X_train.shape)\n",
        "print(\"testing input shape: \", X_test.shape)"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "training input shape:  torch.Size([1000, 24, 1])\n",
            "testing input shape:  torch.Size([300, 24, 1])\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q9HJ-wXAij5k"
      },
      "source": [
        "train_loader = DataLoader(dataset=X_train, \n",
        "                          batch_size=batch_size, \n",
        "                          shuffle=True)\n",
        "\n",
        "test_loader = DataLoader(dataset=X_test, \n",
        "                         batch_size=batch_size, \n",
        "                         shuffle=False)"
      ],
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1wWeovtKij5m",
        "outputId": "17f10f67-406a-4caa-96c9-5f5652a754f4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "##########################\n",
        "### SETTINGS\n",
        "##########################\n",
        "\n",
        "# Device\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "print('Device:', device)\n",
        "\n",
        "# Hyperparameters\n",
        "learning_rate = 0.001\n",
        "num_epochs = 100"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Device: cuda:0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zUlXYUgzij5p"
      },
      "source": [
        "Autoencoder model is created here. The model will take input and return output of the same shape."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SxmKz2obij5r"
      },
      "source": [
        "class ConvolutionalAutoencoder(torch.nn.Module):\n",
        "\n",
        "    def __init__(self):\n",
        "        super(ConvolutionalAutoencoder, self).__init__()\n",
        "        \n",
        "\n",
        "        ### ENCODER\n",
        "        self.conv_1 = torch.nn.Conv1d(in_channels=1,\n",
        "                                      out_channels=32,\n",
        "                                      kernel_size=7,\n",
        "                                      stride=2,\n",
        "                                      padding=1)        \n",
        "        self.dropout_1 = torch.nn.Dropout(p=0.2)                                 \n",
        "        self.conv_2 = torch.nn.Conv1d(in_channels=32,\n",
        "                                      out_channels=16,\n",
        "                                      kernel_size=7,\n",
        "                                      stride=2,\n",
        "                                      padding=1)                 \n",
        "        self.dropout_2 = torch.nn.Dropout(p=0.2)       \n",
        "\n",
        "        ### DECODER                       \n",
        "        self.deconv_1 = torch.nn.ConvTranspose1d(in_channels=16,\n",
        "                                                 out_channels=32,\n",
        "                                                 kernel_size=7,\n",
        "                                                 stride=2,\n",
        "                                                 padding=1)\n",
        "        self.dropout_3 = torch.nn.Dropout(p=0.2)  \n",
        "        self.deconv_2 = torch.nn.ConvTranspose1d(in_channels=32,\n",
        "                                                 out_channels=32,\n",
        "                                                 kernel_size=7,\n",
        "                                                 stride=2,\n",
        "                                                 padding=1)\n",
        "        self.deconv_3 = torch.nn.ConvTranspose1d(in_channels=32,\n",
        "                                                 out_channels=1,\n",
        "                                                 kernel_size=7,\n",
        "                                                 stride=1,\n",
        "                                                 padding=1)\n",
        "        \n",
        "        \n",
        "    def forward(self, x):\n",
        "        ### Input: (batch_size, seq_len, input_size)\n",
        "        ### ENCODER\n",
        "        batch_size = x.size()[0]\n",
        "        seq_length = x.size()[1]\n",
        "        x = x.view(batch_size,1,seq_length)\n",
        "        x = self.conv_1(x)\n",
        "        x = F.relu(x)\n",
        "        x = self.dropout_1(x)\n",
        "        x = self.conv_2(x)\n",
        "        x = F.relu(x)\n",
        "        x = self.dropout_2(x)\n",
        "        \n",
        "        ### DECODER\n",
        "        x = self.deconv_1(x)\n",
        "        x = F.relu(x)\n",
        "        x = self.dropout_3(x)\n",
        "        x = self.deconv_2(x)\n",
        "        x = F.relu(x)\n",
        "        x = self.deconv_3(x)\n",
        "        x = x[:,:,:seq_length]\n",
        "        x = x.view(batch_size,seq_length,1)\n",
        "        return x\n",
        "\n",
        "    \n",
        "# torch.manual_seed(random_seed)\n",
        "model = ConvolutionalAutoencoder()\n",
        "model = model.to(device)\n",
        "\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)  "
      ],
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hG1cOzCCij5s",
        "outputId": "a4b89fe6-a9b3-4e1f-e76a-ac9eae03eb21",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "model"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "ConvolutionalAutoencoder(\n",
              "  (conv_1): Conv1d(1, 32, kernel_size=(7,), stride=(2,), padding=(1,))\n",
              "  (dropout_1): Dropout(p=0.2, inplace=False)\n",
              "  (conv_2): Conv1d(32, 16, kernel_size=(7,), stride=(2,), padding=(1,))\n",
              "  (dropout_2): Dropout(p=0.2, inplace=False)\n",
              "  (deconv_1): ConvTranspose1d(16, 32, kernel_size=(7,), stride=(2,), padding=(1,))\n",
              "  (dropout_3): Dropout(p=0.2, inplace=False)\n",
              "  (deconv_2): ConvTranspose1d(32, 32, kernel_size=(7,), stride=(2,), padding=(1,))\n",
              "  (deconv_3): ConvTranspose1d(32, 1, kernel_size=(7,), stride=(1,), padding=(1,))\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vKhTMUaBij5v"
      },
      "source": [
        "Train the model."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K1IoG5KOij5v",
        "outputId": "b5c186c1-b5e1-4e12-8177-9de2db68cf25",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "start_time = time.time()\n",
        "for epoch in range(num_epochs):\n",
        "    for batch_idx, (features) in enumerate(train_loader):\n",
        "        \n",
        "        # don't need labels, only the images (features)\n",
        "        features = features.to(device)\n",
        "\n",
        "        ### FORWARD AND BACK PROP\n",
        "        logits = model(features)\n",
        "        cost = F.mse_loss(logits, features)\n",
        "        optimizer.zero_grad()\n",
        "        \n",
        "        cost.backward()\n",
        "        \n",
        "        ### UPDATE MODEL PARAMETERS\n",
        "        optimizer.step()\n",
        "        \n",
        "        ### LOGGING\n",
        "        if not batch_idx % 5 and not epoch % 5:\n",
        "            print ('Epoch: %03d/%03d | Batch %03d/%03d | Cost: %.4f' \n",
        "                   %(epoch+1, num_epochs, batch_idx, \n",
        "                     len(X_train)//batch_size, cost))\n",
        "            \n",
        "    print('Time elapsed: %.2f min' % ((time.time() - start_time)/60))\n",
        "    \n",
        "# print('Total Training Time: %.2f min' % ((time.time() - start_time)/60))"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch: 001/100 | Batch 000/007 | Cost: 0.7682\n",
            "Epoch: 001/100 | Batch 005/007 | Cost: 0.8018\n",
            "Time elapsed: 0.00 min\n",
            "Time elapsed: 0.00 min\n",
            "Time elapsed: 0.00 min\n",
            "Time elapsed: 0.00 min\n",
            "Time elapsed: 0.00 min\n",
            "Epoch: 006/100 | Batch 000/007 | Cost: 0.3898\n",
            "Epoch: 006/100 | Batch 005/007 | Cost: 0.2225\n",
            "Time elapsed: 0.00 min\n",
            "Time elapsed: 0.00 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Epoch: 011/100 | Batch 000/007 | Cost: 0.3602\n",
            "Epoch: 011/100 | Batch 005/007 | Cost: 0.1763\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Epoch: 016/100 | Batch 000/007 | Cost: 0.3334\n",
            "Epoch: 016/100 | Batch 005/007 | Cost: 0.3780\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Epoch: 021/100 | Batch 000/007 | Cost: 0.2987\n",
            "Epoch: 021/100 | Batch 005/007 | Cost: 0.1593\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.01 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Epoch: 026/100 | Batch 000/007 | Cost: 0.0993\n",
            "Epoch: 026/100 | Batch 005/007 | Cost: 0.2322\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Epoch: 031/100 | Batch 000/007 | Cost: 0.2238\n",
            "Epoch: 031/100 | Batch 005/007 | Cost: 0.0400\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Epoch: 036/100 | Batch 000/007 | Cost: 0.2240\n",
            "Epoch: 036/100 | Batch 005/007 | Cost: 0.1754\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.02 min\n",
            "Time elapsed: 0.03 min\n",
            "Epoch: 041/100 | Batch 000/007 | Cost: 0.1168\n",
            "Epoch: 041/100 | Batch 005/007 | Cost: 0.0642\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Epoch: 046/100 | Batch 000/007 | Cost: 0.1586\n",
            "Epoch: 046/100 | Batch 005/007 | Cost: 0.2027\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Epoch: 051/100 | Batch 000/007 | Cost: 0.0705\n",
            "Epoch: 051/100 | Batch 005/007 | Cost: 0.0530\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Time elapsed: 0.03 min\n",
            "Epoch: 056/100 | Batch 000/007 | Cost: 0.0791\n",
            "Epoch: 056/100 | Batch 005/007 | Cost: 0.0858\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Epoch: 061/100 | Batch 000/007 | Cost: 0.1232\n",
            "Epoch: 061/100 | Batch 005/007 | Cost: 0.1684\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Epoch: 066/100 | Batch 000/007 | Cost: 0.1947\n",
            "Epoch: 066/100 | Batch 005/007 | Cost: 0.0378\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Epoch: 071/100 | Batch 000/007 | Cost: 0.1217\n",
            "Epoch: 071/100 | Batch 005/007 | Cost: 0.0398\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.04 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Epoch: 076/100 | Batch 000/007 | Cost: 0.0932\n",
            "Epoch: 076/100 | Batch 005/007 | Cost: 0.1349\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Epoch: 081/100 | Batch 000/007 | Cost: 0.1201\n",
            "Epoch: 081/100 | Batch 005/007 | Cost: 0.0889\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Epoch: 086/100 | Batch 000/007 | Cost: 0.0837\n",
            "Epoch: 086/100 | Batch 005/007 | Cost: 0.0510\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.05 min\n",
            "Time elapsed: 0.06 min\n",
            "Time elapsed: 0.06 min\n",
            "Epoch: 091/100 | Batch 000/007 | Cost: 0.0655\n",
            "Epoch: 091/100 | Batch 005/007 | Cost: 0.0918\n",
            "Time elapsed: 0.06 min\n",
            "Time elapsed: 0.06 min\n",
            "Time elapsed: 0.06 min\n",
            "Time elapsed: 0.06 min\n",
            "Time elapsed: 0.06 min\n",
            "Epoch: 096/100 | Batch 000/007 | Cost: 0.0445\n",
            "Epoch: 096/100 | Batch 005/007 | Cost: 0.1958\n",
            "Time elapsed: 0.06 min\n",
            "Time elapsed: 0.06 min\n",
            "Time elapsed: 0.06 min\n",
            "Time elapsed: 0.06 min\n",
            "Time elapsed: 0.06 min\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4_fp-js6ij5x"
      },
      "source": [
        "### Detecting anomalies\n",
        "Anomalies is detected by comparing input and output and determine those with mean absolute error that is higher than threshold.\n",
        "\n",
        "If the MEA loss is greater than  threshold value then it can be infer that the model is seeing a pattern that it isn't familiar with which is a potential anomaly. In this case, we used maximum error from training set."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F6lzjYTrij5x"
      },
      "source": [
        "x_train_pred = model(X_train)\n",
        "x_train_pred = x_train_pred.detach().cpu().numpy() "
      ],
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zawGDzANij50",
        "outputId": "72da3428-1b79-4354-f7ef-266e104a6a1e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        }
      },
      "source": [
        "train_mae_loss = np.mean(np.abs(x_train_pred - train), axis=1)\n",
        "\n",
        "plt.hist(train_mae_loss, bins=50)\n",
        "plt.xlabel(\"Train MAE loss\")\n",
        "plt.ylabel(\"No of samples\")\n",
        "plt.show()\n",
        "\n",
        "# Get reconstruction loss threshold.\n",
        "threshold = np.max(train_mae_loss)\n",
        "print(\"Reconstruction error threshold: \", threshold)"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Reconstruction error threshold:  1.0729535507477488\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AA39w4ITij51"
      },
      "source": [
        "We now validate it by looking at test set for index of time that has high MAE."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D4-Y8zxuij52",
        "outputId": "2b1f4d8b-adc7-4d80-ca94-746c3b3f4b6c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 315
        }
      },
      "source": [
        "x_test_pred = model(X_test)\n",
        "x_test_pred = x_test_pred.detach().cpu().numpy() \n",
        "test_mae_loss = np.mean(np.abs(x_test_pred - test), axis=1)\n",
        "test_mae_loss = test_mae_loss.reshape((-1))\n",
        "\n",
        "plt.hist(test_mae_loss, bins=50)\n",
        "plt.xlabel(\"test MAE loss\")\n",
        "plt.ylabel(\"No of samples\")\n",
        "plt.show()\n",
        "\n",
        "# Detect all the samples which are anomalies.\n",
        "anomalies = (test_mae_loss > threshold).tolist()\n",
        "index = np.where(anomalies)\n",
        "print(\"Number of anomaly samples: \", np.sum(anomalies))\n",
        "print(\"Indices of anomaly samples: \", np.where(anomalies))"
      ],
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Number of anomaly samples:  4\n",
            "Indices of anomaly samples:  (array([105, 112, 119, 124]),)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9vvL-2jXij54"
      },
      "source": [
        "Test_Autoencoder = Autoencoder_Data[1000:]"
      ],
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XGmb92GGij57"
      },
      "source": [
        "anomalous_data_indices = []\n",
        "for data_idx in range(24 - 1, len(Test_Autoencoder) - 24 + 1):\n",
        "    time_series = range(data_idx - 24 + 1, data_idx)\n",
        "    if any([anomalies[j] for j in time_series]):\n",
        "        Test_Autoencoder['outlier'].iloc[data_idx,:] = 1\n",
        "        anomalous_data_indices.append(data_idx)"
      ],
      "execution_count": 77,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y7ZUYcKYRI0I",
        "outputId": "bf88c1f5-5de8-45d6-ec0f-e91ab30de396",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "Test_Autoencoder[Test_Autoencoder['outlier']==1]"
      ],
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>timestamp</th>\n",
              "      <th>key</th>\n",
              "      <th>value</th>\n",
              "      <th>aggregate</th>\n",
              "      <th>total</th>\n",
              "      <th>outlier</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>1106</th>\n",
              "      <td>2015-04-14 01:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>634</td>\n",
              "      <td>1052364</td>\n",
              "      <td>1721</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1107</th>\n",
              "      <td>2015-04-14 02:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1409</td>\n",
              "      <td>1053773</td>\n",
              "      <td>2658</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1108</th>\n",
              "      <td>2015-04-14 03:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1001</td>\n",
              "      <td>1054774</td>\n",
              "      <td>2054</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1109</th>\n",
              "      <td>2015-04-14 04:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1116</td>\n",
              "      <td>1055890</td>\n",
              "      <td>2104</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1110</th>\n",
              "      <td>2015-04-14 05:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>539</td>\n",
              "      <td>1056429</td>\n",
              "      <td>1312</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1111</th>\n",
              "      <td>2015-04-14 06:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>548</td>\n",
              "      <td>1056977</td>\n",
              "      <td>1279</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1112</th>\n",
              "      <td>2015-04-14 07:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>391</td>\n",
              "      <td>1057368</td>\n",
              "      <td>1251</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1113</th>\n",
              "      <td>2015-04-14 08:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>274</td>\n",
              "      <td>1057642</td>\n",
              "      <td>1213</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1114</th>\n",
              "      <td>2015-04-14 09:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>262</td>\n",
              "      <td>1057904</td>\n",
              "      <td>1173</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1115</th>\n",
              "      <td>2015-04-14 10:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>380</td>\n",
              "      <td>1058284</td>\n",
              "      <td>1307</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1116</th>\n",
              "      <td>2015-04-14 11:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>431</td>\n",
              "      <td>1058715</td>\n",
              "      <td>1537</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1117</th>\n",
              "      <td>2015-04-14 12:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>546</td>\n",
              "      <td>1059261</td>\n",
              "      <td>2621</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1118</th>\n",
              "      <td>2015-04-14 13:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>690</td>\n",
              "      <td>1059951</td>\n",
              "      <td>2938</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1119</th>\n",
              "      <td>2015-04-14 14:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>9164</td>\n",
              "      <td>1069115</td>\n",
              "      <td>13646</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1120</th>\n",
              "      <td>2015-04-14 15:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>2203</td>\n",
              "      <td>1071318</td>\n",
              "      <td>4203</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1121</th>\n",
              "      <td>2015-04-14 16:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>2180</td>\n",
              "      <td>1073498</td>\n",
              "      <td>4041</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1122</th>\n",
              "      <td>2015-04-14 17:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1731</td>\n",
              "      <td>1075229</td>\n",
              "      <td>3593</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1123</th>\n",
              "      <td>2015-04-14 18:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1359</td>\n",
              "      <td>1076588</td>\n",
              "      <td>3448</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1124</th>\n",
              "      <td>2015-04-14 19:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1311</td>\n",
              "      <td>1077899</td>\n",
              "      <td>3254</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1125</th>\n",
              "      <td>2015-04-14 20:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1415</td>\n",
              "      <td>1079314</td>\n",
              "      <td>3565</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1126</th>\n",
              "      <td>2015-04-14 21:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>2964</td>\n",
              "      <td>1082278</td>\n",
              "      <td>4722</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1127</th>\n",
              "      <td>2015-04-14 22:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>16628</td>\n",
              "      <td>1098906</td>\n",
              "      <td>18234</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1128</th>\n",
              "      <td>2015-04-14 23:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>68745</td>\n",
              "      <td>1167651</td>\n",
              "      <td>70052</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1129</th>\n",
              "      <td>2015-04-15 00:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>3857</td>\n",
              "      <td>1171508</td>\n",
              "      <td>4996</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1130</th>\n",
              "      <td>2015-04-15 01:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1879</td>\n",
              "      <td>1173387</td>\n",
              "      <td>3075</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1131</th>\n",
              "      <td>2015-04-15 02:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>2008</td>\n",
              "      <td>1175395</td>\n",
              "      <td>3310</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1132</th>\n",
              "      <td>2015-04-15 03:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1294</td>\n",
              "      <td>1176689</td>\n",
              "      <td>2401</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1133</th>\n",
              "      <td>2015-04-15 04:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1227</td>\n",
              "      <td>1177916</td>\n",
              "      <td>1994</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1134</th>\n",
              "      <td>2015-04-15 05:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1522</td>\n",
              "      <td>1179438</td>\n",
              "      <td>2457</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1135</th>\n",
              "      <td>2015-04-15 06:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1282</td>\n",
              "      <td>1180720</td>\n",
              "      <td>2314</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1136</th>\n",
              "      <td>2015-04-15 07:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>521</td>\n",
              "      <td>1181241</td>\n",
              "      <td>1357</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1137</th>\n",
              "      <td>2015-04-15 08:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>488</td>\n",
              "      <td>1181729</td>\n",
              "      <td>1527</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1138</th>\n",
              "      <td>2015-04-15 09:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>428</td>\n",
              "      <td>1182157</td>\n",
              "      <td>1334</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1139</th>\n",
              "      <td>2015-04-15 10:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>492</td>\n",
              "      <td>1182649</td>\n",
              "      <td>2062</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1140</th>\n",
              "      <td>2015-04-15 11:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>619</td>\n",
              "      <td>1183268</td>\n",
              "      <td>2360</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1141</th>\n",
              "      <td>2015-04-15 12:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>708</td>\n",
              "      <td>1183976</td>\n",
              "      <td>2504</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1142</th>\n",
              "      <td>2015-04-15 13:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>733</td>\n",
              "      <td>1184709</td>\n",
              "      <td>3149</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1143</th>\n",
              "      <td>2015-04-15 14:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>986</td>\n",
              "      <td>1185695</td>\n",
              "      <td>3319</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1144</th>\n",
              "      <td>2015-04-15 15:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1154</td>\n",
              "      <td>1186849</td>\n",
              "      <td>3365</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1145</th>\n",
              "      <td>2015-04-15 16:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>1266</td>\n",
              "      <td>1188115</td>\n",
              "      <td>4144</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1146</th>\n",
              "      <td>2015-04-15 17:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>755</td>\n",
              "      <td>1188870</td>\n",
              "      <td>2716</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1147</th>\n",
              "      <td>2015-04-15 18:00:00</td>\n",
              "      <td>AAPL</td>\n",
              "      <td>952</td>\n",
              "      <td>1189822</td>\n",
              "      <td>2860</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "               timestamp   key  value  aggregate  total  outlier\n",
              "1106 2015-04-14 01:00:00  AAPL    634    1052364   1721      1.0\n",
              "1107 2015-04-14 02:00:00  AAPL   1409    1053773   2658      1.0\n",
              "1108 2015-04-14 03:00:00  AAPL   1001    1054774   2054      1.0\n",
              "1109 2015-04-14 04:00:00  AAPL   1116    1055890   2104      1.0\n",
              "1110 2015-04-14 05:00:00  AAPL    539    1056429   1312      1.0\n",
              "1111 2015-04-14 06:00:00  AAPL    548    1056977   1279      1.0\n",
              "1112 2015-04-14 07:00:00  AAPL    391    1057368   1251      1.0\n",
              "1113 2015-04-14 08:00:00  AAPL    274    1057642   1213      1.0\n",
              "1114 2015-04-14 09:00:00  AAPL    262    1057904   1173      1.0\n",
              "1115 2015-04-14 10:00:00  AAPL    380    1058284   1307      1.0\n",
              "1116 2015-04-14 11:00:00  AAPL    431    1058715   1537      1.0\n",
              "1117 2015-04-14 12:00:00  AAPL    546    1059261   2621      1.0\n",
              "1118 2015-04-14 13:00:00  AAPL    690    1059951   2938      1.0\n",
              "1119 2015-04-14 14:00:00  AAPL   9164    1069115  13646      1.0\n",
              "1120 2015-04-14 15:00:00  AAPL   2203    1071318   4203      1.0\n",
              "1121 2015-04-14 16:00:00  AAPL   2180    1073498   4041      1.0\n",
              "1122 2015-04-14 17:00:00  AAPL   1731    1075229   3593      1.0\n",
              "1123 2015-04-14 18:00:00  AAPL   1359    1076588   3448      1.0\n",
              "1124 2015-04-14 19:00:00  AAPL   1311    1077899   3254      1.0\n",
              "1125 2015-04-14 20:00:00  AAPL   1415    1079314   3565      1.0\n",
              "1126 2015-04-14 21:00:00  AAPL   2964    1082278   4722      1.0\n",
              "1127 2015-04-14 22:00:00  AAPL  16628    1098906  18234      1.0\n",
              "1128 2015-04-14 23:00:00  AAPL  68745    1167651  70052      1.0\n",
              "1129 2015-04-15 00:00:00  AAPL   3857    1171508   4996      1.0\n",
              "1130 2015-04-15 01:00:00  AAPL   1879    1173387   3075      1.0\n",
              "1131 2015-04-15 02:00:00  AAPL   2008    1175395   3310      1.0\n",
              "1132 2015-04-15 03:00:00  AAPL   1294    1176689   2401      1.0\n",
              "1133 2015-04-15 04:00:00  AAPL   1227    1177916   1994      1.0\n",
              "1134 2015-04-15 05:00:00  AAPL   1522    1179438   2457      1.0\n",
              "1135 2015-04-15 06:00:00  AAPL   1282    1180720   2314      1.0\n",
              "1136 2015-04-15 07:00:00  AAPL    521    1181241   1357      1.0\n",
              "1137 2015-04-15 08:00:00  AAPL    488    1181729   1527      1.0\n",
              "1138 2015-04-15 09:00:00  AAPL    428    1182157   1334      1.0\n",
              "1139 2015-04-15 10:00:00  AAPL    492    1182649   2062      1.0\n",
              "1140 2015-04-15 11:00:00  AAPL    619    1183268   2360      1.0\n",
              "1141 2015-04-15 12:00:00  AAPL    708    1183976   2504      1.0\n",
              "1142 2015-04-15 13:00:00  AAPL    733    1184709   3149      1.0\n",
              "1143 2015-04-15 14:00:00  AAPL    986    1185695   3319      1.0\n",
              "1144 2015-04-15 15:00:00  AAPL   1154    1186849   3365      1.0\n",
              "1145 2015-04-15 16:00:00  AAPL   1266    1188115   4144      1.0\n",
              "1146 2015-04-15 17:00:00  AAPL    755    1188870   2716      1.0\n",
              "1147 2015-04-15 18:00:00  AAPL    952    1189822   2860      1.0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 80
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PXdTeUWyij5-",
        "outputId": "c6b5c460-c99c-4f65-b2fa-aba1c836e202",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 254
        }
      },
      "source": [
        "df_subset = Test_Autoencoder.iloc[anomalous_data_indices, :]\n",
        "plt.subplots_adjust(bottom=0.2)\n",
        "plt.xticks(rotation=25)\n",
        "ax = plt.gca()\n",
        "\n",
        "values = pd.Series(Test_Autoencoder[\"value\"])\n",
        "plt.plot(values, label=\"test data\")\n",
        "values = pd.Series(df_subset[\"value\"])\n",
        "plt.plot(values, label=\"anomalies\", color=\"r\")\n",
        "\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}