CovAP.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# COVID-19 Analysis Platform"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 430,
   "metadata": {},
   "outputs": [],
   "source": [
    "from jupyter_dash import JupyterDash"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 431,
   "metadata": {},
   "outputs": [],
   "source": [
    "import dash\n",
    "import dash_core_components as dcc\n",
    "import dash_html_components as html\n",
    "import pandas as pd\n",
    "import dash_bootstrap_components as dbc\n",
    "import dash_table as dt\n",
    "from dash.dependencies import Input, Output\n",
    "from statsmodels.tsa.arima_model import ARIMA\n",
    "from pmdarima import auto_arima\n",
    "import warnings\n",
    "from fbprophet import Prophet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 432,
   "metadata": {},
   "outputs": [],
   "source": [
    "debug = pd.read_json(\"properties.json\", orient=\"index\").debug.value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 433,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"owid-covid-data.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 434,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "countries = df.location.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 435,
   "metadata": {},
   "outputs": [],
   "source": [
    "world_code = \"OWID_WRL\"\n",
    "\n",
    "null_fill_columns = {'new_cases': 0, 'new_deaths': 0, 'total_cases': 0, 'total_deaths': 0,\n",
    "                     'total_cases_per_million': 0, 'total_deaths_per_million' : 0, \n",
    "                     'new_cases_smoothed': 0, 'new_cases_smoothed_per_million': 0, 'new_deaths_smoothed': 0, \n",
    "                     'new_deaths_smoothed_per_million': 0}\n",
    "\n",
    "numeric_columns = ['new_cases', 'new_deaths', 'total_cases', 'total_deaths', 'total_cases_per_million',\n",
    "                'total_deaths_per_million']\n",
    "\n",
    "df = df.fillna(null_fill_columns)\n",
    "\n",
    "dfmain = df[df.groupby(['location']).date.transform('max') == df.date]\n",
    "dfmain = dfmain.fillna(null_fill_columns)\n",
    "\n",
    "dft = dfmain[['location', 'continent', 'total_cases', 'total_cases_per_million',\n",
    "                'total_deaths', 'total_deaths_per_million']]\n",
    "\n",
    "dfmain_no_world = dfmain[~dfmain.iso_code.isin(['OWID_WRL'])] # Remove world row for map\n",
    "dfm = dfmain_no_world[['iso_code', 'location', 'total_cases_per_million', 'total_deaths_per_million']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 436,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Alcohol sales data\n",
    "alcohol_sales = pd.read_csv(\"Statistics on alcohol sales, by type of beverage, contents and quarter.csv\", delimiter=\";\")\n",
    "alcohol_sales = alcohol_sales.set_index(\"type of beverage\")\n",
    "alcohol_sales = alcohol_sales.rename(columns={\n",
    " \"Sales of alcoholic beverages per capita 15 years and over, as sold (litres) 2019K1\": \"Q1 2019 Sales\",\n",
    " \"Sales of alcoholic beverages per capita 15 years and over, as sold (litres) 2019K2\": \"Q2 2019 Sales\",\n",
    " \"Sales of alcoholic beverages per capita 15 years and over, as sold (litres) 2019K3\": \"Q3 2019 Sales\",\n",
    " \"Sales of alcoholic beverages per capita 15 years and over, as sold (litres) 2019K4\": \"Q4 2019 Sales\",\n",
    " \"Sales of alcoholic beverages per capita 15 years and over, as sold (litres) 2020K1\": \"Q1 2020 Sales\",\n",
    " \"Sales of alcoholic beverages per capita 15 years and over, as sold (litres) 2020K2\": \"Q2 2020 Sales\",\n",
    " \"Sales of alcoholic beverages per capita 15 years (litres) and over, pure alcohol 2019K1\": \"Q1 2019 Sales pure alcohol\",\n",
    " \"Sales of alcoholic beverages per capita 15 years (litres) and over, pure alcohol 2019K2\": \"Q2 2019 Sales pure alcohol\",\n",
    " \"Sales of alcoholic beverages per capita 15 years (litres) and over, pure alcohol 2019K3\": \"Q3 2019 Sales pure alcohol\",\n",
    " \"Sales of alcoholic beverages per capita 15 years (litres) and over, pure alcohol 2019K4\": \"Q4 2019 Sales pure alcohol\",\n",
    " \"Sales of alcoholic beverages per capita 15 years (litres) and over, pure alcohol 2020K1\": \"Q1 2020 Sales pure alcohol\",\n",
    " \"Sales of alcoholic beverages per capita 15 years (litres) and over, pure alcohol 2020K2\": \"Q2 2020 Sales pure alcohol\",  \n",
    "})\n",
    "\n",
    "alcohol_sales_all = alcohol_sales[[\"Q1 2019 Sales\", \"Q2 2019 Sales\", \"Q3 2019 Sales\", \"Q4 2019 Sales\",\n",
    "                                   \"Q1 2020 Sales\", \"Q2 2020 Sales\"]]\n",
    "alcohol_sales_all = alcohol_sales_all.T\n",
    "alcohol_sales_all.columns.name = None\n",
    "new_indices = [i.replace(\"Sales\", \"\").strip() for i in alcohol_sales_all.index]\n",
    "alcohol_sales_all = alcohol_sales_all.rename(index=dict(zip(alcohol_sales_all.index,new_indices)))\n",
    "\n",
    "alcohol_sales_pure_alcohol = alcohol_sales[[\"Q1 2019 Sales pure alcohol\", \"Q2 2019 Sales pure alcohol\",\n",
    "                                   \"Q3 2019 Sales pure alcohol\", \"Q4 2019 Sales pure alcohol\", \n",
    "                                   \"Q1 2020 Sales pure alcohol\", \"Q2 2020 Sales pure alcohol\"]]\n",
    "alcohol_sales_pure_alcohol = alcohol_sales_pure_alcohol.T\n",
    "alcohol_sales_pure_alcohol.columns.name = None\n",
    "alcohol_sales_pure_alcohol = alcohol_sales_pure_alcohol.rename(\n",
    "    index=dict(zip(alcohol_sales_pure_alcohol.index,new_indices)))\n",
    "\n",
    "# To drop total alcohol sales\n",
    "alcohol_sales = alcohol_sales.drop(alcohol_sales.index[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 437,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Trips data\n",
    "trips = pd.read_csv(\"Trips (million), by type of destination, type of trip, contents and quarter.csv\", delimiter=\";\")\n",
    "grouped_trips = trips.groupby([\"type of destination\", \"type of trip\"]).sum()\n",
    "grouped_trips_transposed = grouped_trips.T\n",
    "new_indices_latest = new_indices.append(\"Q3 2020\")\n",
    "trips_grouped = grouped_trips_transposed.rename(index=dict(zip(grouped_trips_transposed.index,new_indices)))\n",
    "trips = trips.rename(columns=dict(zip(trips.columns[2:],new_indices)))\n",
    "domestic_trips = trips[trips[\"type of trip\"] == \"Total trips domestic\"]\n",
    "\n",
    "# To drop all trips\n",
    "domestic_trips = domestic_trips.drop(domestic_trips.index[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 438,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predictions data\n",
    "predictions = pd.read_csv(\"predictions.csv\")\n",
    "# predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 439,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "def getFormat(column):\n",
    "    if(column in numeric_columns):\n",
    "        if(\"per_million\" in column):\n",
    "            return {'specifier': ',.2f'}\n",
    "        else:\n",
    "            return {'specifier': ',.0f'}\n",
    "    return None\n",
    "\n",
    "    \n",
    "def getType(column):\n",
    "    if(column in numeric_columns):\n",
    "        return \"numeric\"\n",
    "    return \"text\"\n",
    "\n",
    "\n",
    "def getTitleText(text):\n",
    "    return text.title().replace('_', ' ').replace(\"Gdp\", \"GDP\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 440,
   "metadata": {},
   "outputs": [],
   "source": [
    "def replace_negatives(series):\n",
    "    return series.mask(series.lt(0), 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 441,
   "metadata": {},
   "outputs": [],
   "source": [
    "header = html.H1(\"CovAP dashboard\")\n",
    "last_updated = html.H6(\"Last updated on: \"+df.date.max(),className=\"text-right\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 442,
   "metadata": {},
   "outputs": [],
   "source": [
    "comparison_options = ['new_cases_smoothed', 'new_cases_smoothed_per_million', 'new_deaths_smoothed',\n",
    "                      'new_deaths_smoothed_per_million']\n",
    "\n",
    "location_comparison_div = html.Div([\n",
    "    html.Div(html.Div(html.H4(\"Comparison of COVID-19 outbreak by location\"), className=\"col p-0\"),\n",
    "             className=\"row mt-2\"),\n",
    "    \n",
    "    html.Div([\n",
    "        html.Div(html.Label(\"Select one or more locations\"), className=\"col-auto p-0\"),\n",
    "        \n",
    "        html.Div(dcc.Dropdown(\n",
    "            id='multi-country-input',\n",
    "            options=[{'label': i, 'value': i} for i in countries],\n",
    "            value=['Norway', 'Sweden', 'Denmark', 'Finland'],\n",
    "            clearable=False,\n",
    "            multi=True,\n",
    "        ), className=\"col-4\"),\n",
    "        \n",
    "        html.Div(html.Label(\"Select a category\"), className=\"col-auto p-0\"),\n",
    "        \n",
    "        html.Div(\n",
    "            dcc.Dropdown(\n",
    "                id=\"comparison-type\", \n",
    "                options=[{'label': getTitleText(c), 'value': c} for c in comparison_options],\n",
    "                value = comparison_options[0],\n",
    "                clearable=False,\n",
    "            ), className=\"col-4\"),\n",
    "        \n",
    "    ], className=\"row mb-2\"),\n",
    "    \n",
    "    html.Div(html.Div(dcc.Graph(id='comparison-plot'), className=\"col border border-dark\"),\n",
    "             className=\"row align-items-center\"),\n",
    "    \n",
    "], className=\"container-fluid\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 443,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Metric Analysis tab content\n",
    "dot_measures = [\"population\", \"population_density\", \"gdp_per_capita\", \"extreme_poverty\", \"cardiovasc_death_rate\",\n",
    "                \"diabetes_prevalence\", \"life_expectancy\", \"human_development_index\", \"median_age\", \"stringency_index\"]\n",
    "\n",
    "dot_measures.sort()\n",
    "\n",
    "default_analysis_type = dot_measures[2]\n",
    "\n",
    "analysis_locations = dfmain_no_world[~dfmain_no_world[default_analysis_type].isna()].location\n",
    "analysis_locations_list = analysis_locations.tolist()\n",
    "first_location = \"\" if len(analysis_locations_list) == 0 else analysis_locations_list[0]\n",
    "default_location = \"Norway\" if \"Norway\" in analysis_locations_list else first_location\n",
    "\n",
    "dot_div = html.Div([\n",
    "    html.Div(html.Div(html.H4(\"Analysis of COVID-19 prevalence with location wise metrics\"), className=\"col p-0\"),\n",
    "             className=\"row mt-2\"),\n",
    "        \n",
    "    html.Div([\n",
    "            html.Div(html.Label(\"Select a metric\"), className=\"col-auto p-0\"),\n",
    "            html.Div(dcc.Dropdown(\n",
    "                id='dot-metric',\n",
    "                options=[\n",
    "                    {'label': getTitleText(prop), 'value': prop} for prop in dot_measures\n",
    "                ],\n",
    "                value=\"gdp_per_capita\",\n",
    "            ), className=\"col-4\"),\n",
    "            \n",
    "            html.Div(html.Label(\"Select a location to highlight\"), className=\"col-auto p-0\"),\n",
    "        \n",
    "            html.Div(dcc.Dropdown(\n",
    "                id='hightlight-input',\n",
    "                options=[\n",
    "                    {'label': location, 'value': location} for location in analysis_locations\n",
    "                ],\n",
    "                value=default_location,\n",
    "            ), className=\"col-4\"),\n",
    "        \n",
    "    ], className=\"row mb-2\"),\n",
    "    \n",
    "    html.Div([\n",
    "        html.Div(dcc.Graph(id='total-dot-plot'), className=\"col border border-dark\"),\n",
    "        \n",
    "        html.Div(dcc.Graph(id='dot-plot'), className=\"col border-top border-right border-bottom border-dark\"),\n",
    "        \n",
    "    ], className=\"row align-items-center\"),\n",
    "    \n",
    "], className=\"container-fluid\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 444,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Daily stats tab content\n",
    "daily_stats_div = html.Div([\n",
    "    \n",
    "    html.Div(html.Div(html.H4(\"Daily COVID-19 statistics by location\"), className=\"col-auto p-0\"), className=\"row mt-2\"),\n",
    "    \n",
    "    html.Div([\n",
    "        html.Div(html.Label(\"Select a location\"), className=\"col-auto p-0\"),\n",
    "        \n",
    "        html.Div(dcc.Dropdown(\n",
    "            id='country-input',\n",
    "            options=[{'label': i, 'value': i} for i in countries],\n",
    "            value='Norway',\n",
    "            clearable=False,\n",
    "    ), className=\"col-4\"),\n",
    "    ], className=\"row mb-2\"),\n",
    "    \n",
    "    html.Div([\n",
    "        \n",
    "        html.Div(dcc.Graph(id='daily-cases'), className=\"col border border-dark\"),\n",
    "        \n",
    "        html.Div(dcc.Graph(id='daily-deaths'), className=\"col border-top border-right border-bottom border-dark\"),\n",
    "        \n",
    "        html.Div(dcc.Graph(id='daily-tests'), className=\"col border-top border-right border-bottom border-dark\"),\n",
    "        \n",
    "    ], className=\"row align-items-center\"),\n",
    "    \n",
    "], className=\"container-fluid\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 445,
   "metadata": {},
   "outputs": [],
   "source": [
    "world_table = dt.DataTable(\n",
    "    id=\"world-table\",\n",
    "    data=dft.to_dict('records'),\n",
    "    columns=[{'id': c, 'name': getTitleText(c), \"type\": getType(c), 'format': getFormat(c)} for c in dft.columns],\n",
    "    sort_by=[{\"column_id\": 'total_deaths', \"direction\": \"desc\"}],\n",
    "    page_size=12,\n",
    "    sort_action='native',\n",
    "    filter_action=\"native\",\n",
    "    style_cell={\n",
    "            'whiteSpace': 'normal',\n",
    "            'height': 'auto',\n",
    "            'width': '15%',\n",
    "            'font-family': \"Helvetica Neue, Helvetica, Arial, sans-serif\",\n",
    "    },\n",
    "    style_cell_conditional=[\n",
    "        {'if': {'column_id': 'location'},\n",
    "         'width': '22%'},\n",
    "        {'if': {'column_id': 'continent'},\n",
    "         'width': '18%'},\n",
    "    ],\n",
    "    style_data_conditional=[\n",
    "        {\n",
    "            'if': {'row_index': 'odd'},\n",
    "            'backgroundColor': 'rgb(248, 248, 248)'\n",
    "        },\n",
    "        {\n",
    "        'if': {\n",
    "            'filter_query': '{location} = \"World\"'\n",
    "        },\n",
    "        'backgroundColor': 'grey',\n",
    "        'color': 'white'\n",
    "        },\n",
    "        {\n",
    "        'if': {\n",
    "            'column_type': 'text'  # 'text' | 'any' | 'datetime' | 'numeric'\n",
    "        },\n",
    "        'textAlign': 'left'\n",
    "        },\n",
    "    ],\n",
    "    style_header={\n",
    "        'backgroundColor': 'rgb(230, 230, 230)',\n",
    "        'fontWeight': 'bold',\n",
    "        'textAlign': 'center'\n",
    "    },\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 446,
   "metadata": {},
   "outputs": [],
   "source": [
    "map_details_div = html.Div([\n",
    "    html.Div(\n",
    "        html.Div(html.H4(id='click-data'), className=\"col d-flex align-items-center justify-content-center\"),\n",
    "        className=\"row\"),\n",
    "    \n",
    "    html.Div([\n",
    "\n",
    "        html.Div([\n",
    "            html.H5(\"Cases\"),\n",
    "        ], className=\"col d-flex align-items-center justify-content-center text-center p-0 border-right border-dark\"),\n",
    "\n",
    "        html.Div([\n",
    "            html.H5(\"Deaths\"), \n",
    "        ], className=\"col d-flex align-items-center justify-content-center text-center p-0\"),\n",
    "\n",
    "    ], className=\"row border-top border-left border-right border-dark\"),\n",
    "\n",
    "    html.Div([\n",
    "\n",
    "        html.Div([\n",
    "            html.H6(\"Total\"),\n",
    "        ], className=\"col d-flex align-items-center justify-content-center text-center p-0 border-right border-dark\"),\n",
    "\n",
    "        html.Div([\n",
    "            html.H6(\"Per Million\"),\n",
    "        ], className=\"col d-flex align-items-center justify-content-center text-center p-0 border-right border-dark\"),\n",
    "\n",
    "        html.Div([\n",
    "            html.H6(\"Total\"),\n",
    "        ], className=\"col d-flex align-items-center justify-content-center text-center p-0 border-right border-dark\"),\n",
    "\n",
    "        html.Div([\n",
    "            html.H6(\"Per Million\"), \n",
    "        ], className=\"col d-flex align-items-center justify-content-center text-center p-0\"),\n",
    "\n",
    "    ], className=\"row border-top border-left border-right border-dark\"),\n",
    "    html.Div([\n",
    "\n",
    "        html.Div(html.Span(id=\"total-cases\"),\n",
    "                 className=\"col d-flex align-items-center justify-content-center border-right border-dark\"),\n",
    "\n",
    "        html.Div(html.Span(id=\"total-cases-pm\"), \n",
    "                 className=\"col d-flex align-items-center justify-content-center border-right border-dark\"),\n",
    "\n",
    "        html.Div(html.Span(id=\"total-deaths\"),\n",
    "                 className=\"col d-flex align-items-center justify-content-center border-right border-dark\"),\n",
    "\n",
    "        html.Div(html.Span(id=\"total-deaths-pm\"), \n",
    "                 className=\"col d-flex align-items-center justify-content-center\"),\n",
    "\n",
    "    ], className=\"row border border-dark\"),\n",
    "\n",
    "    html.Div(html.Div(\n",
    "        html.Button(\"Reset\", id='btn-world', className=\"btn btn-light btn-sm\",\n",
    "                    role=\"button\"),\n",
    "        className=\"col d-flex align-items-center justify-content-end p-0 mt-2\"), className=\"row\"),\n",
    "\n",
    "], className=\"container-fluid\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 447,
   "metadata": {},
   "outputs": [],
   "source": [
    "# World data tab content\n",
    "map_options = ['total_deaths_per_million', 'total_cases_per_million']\n",
    "world_data_div = html.Div([\n",
    "    \n",
    "    html.Div(html.Div(html.H4(\"COVID-19 statistics for the world\"), className=\"col-auto p-0\"), className=\"row mt-2\"),\n",
    "    \n",
    "    html.Div([\n",
    "        html.Div(html.Label(\"Select a visual category\"), className=\"col-auto p-0\"),\n",
    "        \n",
    "        html.Div(\n",
    "        dcc.RadioItems(\n",
    "            id='world-type',\n",
    "            options=[\n",
    "                {'label': \"Map\", 'value': \"map\"},\n",
    "                {'label': \"Table\", 'value': \"table\"}\n",
    "            ],\n",
    "            value=\"map\",\n",
    "            labelClassName=\"mr-2\"\n",
    "        ), className=\"col\")\n",
    "        \n",
    "    ], className=\"row\"),\n",
    "    \n",
    "    html.Div([ #row 3\n",
    "            \n",
    "            html.Div([\n",
    "                html.Div([\n",
    "                    \n",
    "                    html.Div([\n",
    "                        \n",
    "                        html.Div(html.Label(\"Select a map type\"), className=\"col-auto p-0\"),\n",
    "                        html.Div(\n",
    "                            dcc.RadioItems(\n",
    "                                id='map-type',\n",
    "                                options=[{'label': getTitleText(i), 'value': i} for i in map_options ],\n",
    "                                value=map_options[0],\n",
    "                                labelClassName=\"mr-2\"\n",
    "                            ), className=\"col\"),\n",
    "                        ],className=\"row mt-2\"),\n",
    "                    \n",
    "                    html.Div([\n",
    "                        \n",
    "                        html.Div(dcc.Graph(id='world-map'), className=\"col-8 p-0\"),\n",
    "                        \n",
    "                        html.Div(map_details_div, className=\"col d-flex align-items-center justify-content-center p-0\"),\n",
    "                        \n",
    "                    ], className=\"row align-items-center\"),\n",
    "                    \n",
    "                    html.Div(\n",
    "                        html.Div(\"Click on a map location to see details on the right.\",\n",
    "                                 className=\"col text-info text-left p-0\"),\n",
    "                    className=\"row\"),\n",
    "                    \n",
    "                ], className=\"container-fluid\"),\n",
    "                \n",
    "            ], className=\"col border border-dark\", id=\"world-map-div\",),  \n",
    "        \n",
    "        html.Div(world_table, className=\"col border border-dark\", id=\"world-table-div\",),   \n",
    "            \n",
    "    ], className=\"row\"),\n",
    "        \n",
    "], className=\"container-fluid\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 448,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predictions tab content\n",
    "prediction_columns = [\"new_cases_smoothed\", \"new_deaths_smoothed\"]\n",
    "predictions_div = html.Div([\n",
    "    \n",
    "    html.Div(html.Div(html.H4(\"COVID-19 predictions for Nordic countries\"), className=\"col-auto p-0\"), className=\"row mt-2\"),\n",
    "    \n",
    "    html.Div([\n",
    "        html.Div(html.Label(\"Select a location\"), className=\"col-auto p-0\"),\n",
    "        html.Div(dcc.Dropdown(\n",
    "            id='prediction-location',\n",
    "            options=[{'label': i, 'value': i} for i in predictions.location.unique()],\n",
    "            value='Norway',\n",
    "            clearable=False,\n",
    "    ), className=\"col-4\"),\n",
    "        html.Div(html.Label(\"Select a category\"), className=\"col-auto p-0\"),\n",
    "        html.Div(\n",
    "            dcc.RadioItems(\n",
    "                id=\"prediction-column\", \n",
    "                options=[{'label': getTitleText(c), 'value': c} for c in prediction_columns],\n",
    "                           value = prediction_columns[0],\n",
    "                          labelClassName=\"mr-2\"), \n",
    "            className=\"col-auto\")\n",
    "    ], className=\"row mb-2\"),\n",
    "    \n",
    "    html.Div([\n",
    "        \n",
    "        html.Div(dcc.Graph(id='prediction-new-cases'), className=\"col border border-dark\"),\n",
    "        \n",
    "        html.Div(dcc.Graph(id='prediction-new-deaths'), className=\"col border-top border-right border-bottom border-dark\"),\n",
    "        \n",
    "    ], className=\"row align-items-center\"),\n",
    "    \n",
    "], className=\"container-fluid\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 449,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Datasets tab content\n",
    "datasets_options = [\"grouped\", \"show_trend\"]\n",
    "datasets_div = html.Div([\n",
    "    html.Div(html.Div(html.H4(\"Impact of COVID-19 pandemic on Norwegian population\"), className=\"col p-0\"),\n",
    "             className=\"row mt-2\"),\n",
    "    html.Div([\n",
    "        html.Div(\n",
    "            html.Div([\n",
    "                html.Div([\n",
    "                    html.Div(html.Label(\"Select a category\"), className=\"col-auto pr-0\"),\n",
    "                    \n",
    "                    html.Div(\n",
    "                        dcc.RadioItems(\n",
    "                            id='alcohol-stats-type',\n",
    "                            options=[{'label': getTitleText(i), 'value': i} for i in datasets_options ],\n",
    "                            value=datasets_options[0],\n",
    "                            labelClassName=\"mr-2\"\n",
    "                        ), className=\"col\"),\n",
    "                    \n",
    "                ], className=\"row mt-2\"),\n",
    "\n",
    "                    html.Div(html.Div(dcc.Graph(id='alcohol-stats'), className=\"col\"),\n",
    "                             className=\"row align-items-center\"),\n",
    "\n",
    "                ], className=\"container-fluid\"), className=\"col p-0 border border-dark\"),\n",
    "\n",
    "        html.Div(\n",
    "            html.Div([\n",
    "                html.Div([\n",
    "                    html.Div(html.Label(\"Select a category\"), className=\"col-auto pr-0\"),\n",
    "                    \n",
    "                    html.Div(\n",
    "                        dcc.RadioItems(\n",
    "                            id='trips-stats-type',\n",
    "                            options=[{'label': getTitleText(i), 'value': i} for i in datasets_options ],\n",
    "                            value=datasets_options[0],\n",
    "                            labelClassName=\"mr-2\"\n",
    "                        ), className=\"col\"),\n",
    "                    \n",
    "                ], className=\"row mt-2\"),\n",
    "\n",
    "                    html.Div(html.Div(dcc.Graph(id='trips-stats'), className=\"col\"), \n",
    "                         className=\"row align-items-center\"),\n",
    "\n",
    "                ], className=\"container-fluid\"), className=\"col p-0 border-top border-right border-bottom border-dark\"),\n",
    "    \n",
    "    ], className=\"row\"),\n",
    "    \n",
    "    html.Div(html.Div([\n",
    "        \"Source: \", html.A(\"Statistics Norway\", href=\"https://www.ssb.no/en/statbank\", target=\"_blank\")\n",
    "    ], className=\"col\"), className=\"row\"),\n",
    "    \n",
    "],className=\"container-fluid\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 450,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "external_stylesheets = ['https://codepen.io/chriddyp/pen/bWLwgP.css']\n",
    "\n",
    "# app = JupyterDash(__name__, external_stylesheets=external_stylesheets)\n",
    "app = JupyterDash(__name__, external_stylesheets=[dbc.themes.BOOTSTRAP])\n",
    "\n",
    "# Create server variable with Flask server object for use with gunicorn\n",
    "server = app.server\n",
    "\n",
    "app.layout = html.Div([\n",
    "    html.Div([\n",
    "        html.Div(className=\"col\"), \n",
    "        \n",
    "        html.Div(header, className=\"col d-flex align-items-center justify-content-center\"), \n",
    "        \n",
    "        html.Div(last_updated, className=\"col d-flex align-items-end justify-content-end\")\n",
    "        \n",
    "    ], className=\"row\"),\n",
    "    \n",
    "    html.Div(html.Div(dcc.Tabs([\n",
    "        dcc.Tab(label='Daily Statistics', children=[\n",
    "            daily_stats_div\n",
    "        ]),\n",
    "        dcc.Tab(label='World Data', children=[\n",
    "            world_data_div\n",
    "        ]),\n",
    "        dcc.Tab(label='Metric Analysis', children=[\n",
    "            dot_div\n",
    "        ]),\n",
    "        dcc.Tab(label='Location Comparison', children=[\n",
    "            location_comparison_div\n",
    "        ]),\n",
    "        dcc.Tab(label='Impact Study', children=[\n",
    "            datasets_div\n",
    "        ]),\n",
    "        dcc.Tab(label='COVID-19 Predictions', children=[\n",
    "            predictions_div\n",
    "        ]),\n",
    "    ]),className=\"col\"),className=\"row\"),\n",
    "], className=\"container-fluid\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 451,
   "metadata": {},
   "outputs": [],
   "source": [
    "# World Type callback\n",
    "@app.callback([\n",
    "    Output('world-map-div', 'style'),\n",
    "    Output('world-table-div', 'style'),\n",
    "], Input('world-type', 'value'))\n",
    "def set_world_type_plot(world_type):\n",
    "    if (world_type == \"map\"):\n",
    "        return {\"display\": \"block\"}, {\"display\": \"none\"}\n",
    "    else:\n",
    "        return {\"display\": \"none\"}, {\"display\": \"block\"}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 452,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_empty_graph(text):\n",
    "    return {\n",
    "        \"layout\": {\n",
    "            \"xaxis\": {\n",
    "                \"visible\": False\n",
    "            },\n",
    "            \"yaxis\": {\n",
    "                \"visible\": False\n",
    "            },\n",
    "            \"annotations\": [\n",
    "                {\n",
    "                    \"text\": \"No \"+ text.lower() +\" data\",\n",
    "                    \"xref\": \"paper\",\n",
    "                    \"yref\": \"paper\",\n",
    "                    \"showarrow\": False,\n",
    "                    \"font\": {\n",
    "                        \"size\": 20\n",
    "                    }\n",
    "                }\n",
    "            ]\n",
    "        }\n",
    "    }\n",
    "\n",
    "def get_daily_graph(x, y, text, color):\n",
    "    if(pd.isna(y.max())):\n",
    "        return get_empty_graph(text)\n",
    "    return  {\n",
    "        'data': [dict(\n",
    "            x=x,\n",
    "            y=y,\n",
    "            type='bar',\n",
    "            marker={\n",
    "                'color': color\n",
    "            },\n",
    "            hovertemplate=\"%{y:,}<br>%{x}<extra></extra>\"\n",
    "        )],\n",
    "        'layout': dict(\n",
    "            margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "            hovermode='closest',\n",
    "            title={\n",
    "            'text': text,\n",
    "            'y':0.9,\n",
    "            'x':0.5,\n",
    "            'xanchor': 'center',\n",
    "            'yanchor': 'top',\n",
    "            'font': dict(size=20),\n",
    "            },\n",
    "            hoverlabel=dict(\n",
    "                bgcolor=\"white\",\n",
    "                font=dict(\n",
    "                    size=16,\n",
    "                    family=\"Rockwell\"\n",
    "                ),\n",
    "            ),\n",
    "        )\n",
    "    }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 453,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Daily stats tab callback\n",
    "@app.callback([\n",
    "    Output('daily-cases', 'figure'),\n",
    "    Output('daily-deaths', 'figure'),\n",
    "    Output('daily-tests', 'figure'),\n",
    "],\n",
    "[\n",
    "    Input('country-input', 'value'),\n",
    "])\n",
    "def update_daily_stats(location):\n",
    "\n",
    "    dff = df[df.location == location]\n",
    "       \n",
    "    return (get_daily_graph(dff.date, dff.new_cases,\"Daily New Cases\", \"crimson\"),\n",
    "            get_daily_graph(dff.date, dff.new_deaths,\"Daily New Deaths\", \"darkslateblue\"),\n",
    "            get_daily_graph(dff.date, dff.new_tests,\"Daily New Tests\", \"deepskyblue\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 454,
   "metadata": {},
   "outputs": [],
   "source": [
    "alcohol_hovertemplate=\"%{y} litres<extra></extra>\"\n",
    "alcohol_title_text = \"Sales of alcoholic beverages per capita\"\n",
    "\n",
    "def get_alcohol_sales_previous_year():\n",
    "    x = alcohol_sales.index\n",
    "    q2_2019 = \"Q2 2019\"\n",
    "    q2_2020 = \"Q2 2020\"\n",
    "    \n",
    "    return {\n",
    "            'data': [dict(\n",
    "                x=x,\n",
    "                y=alcohol_sales[\"Q2 2019 Sales\"],\n",
    "                type='bar',\n",
    "                marker={\n",
    "                    'color': 'orange'\n",
    "                },\n",
    "                name=q2_2019,\n",
    "                hovertemplate=q2_2019 +\"<br>\"+ alcohol_hovertemplate,\n",
    "            ),\n",
    "            dict(\n",
    "                x=x,\n",
    "                y=alcohol_sales[\"Q2 2020 Sales\"],\n",
    "                type='bar',\n",
    "                marker={\n",
    "                    'color': 'red'\n",
    "                },\n",
    "                name=q2_2020,\n",
    "                hovertemplate=q2_2020 +\"<br>\"+ alcohol_hovertemplate,\n",
    "            )],\n",
    "            'layout': dict(\n",
    "                margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "                hovermode='closest',\n",
    "                title={\n",
    "                'text': getTitleText(alcohol_title_text),\n",
    "                'y':0.9,\n",
    "                'x':0.5,\n",
    "                'xanchor': 'center',\n",
    "                'yanchor': 'top',\n",
    "                'font': dict(size=20),\n",
    "                },\n",
    "                hoverlabel=dict(\n",
    "                    bgcolor=\"white\",\n",
    "                    font=dict(\n",
    "                        size=16,\n",
    "                        family=\"Rockwell\"\n",
    "                    ),\n",
    "                ),\n",
    "                legend=dict(\n",
    "                    title=dict(\n",
    "                        text=\"Quarters\",\n",
    "                    )\n",
    "                ),\n",
    "            )\n",
    "        }\n",
    "\n",
    "def get_alcohol_sales_trend():\n",
    "    data = []\n",
    "    \n",
    "    for alcohol in alcohol_sales_all.columns:\n",
    "        sales = alcohol_sales_all[alcohol]\n",
    "        sub_data = dict(\n",
    "            x=sales.index,\n",
    "            y=sales,\n",
    "            type='scatter',\n",
    "            name=alcohol,\n",
    "            hovertemplate=alcohol +\" : \"+ alcohol_hovertemplate,\n",
    "        )\n",
    "        data.append(sub_data)\n",
    "        \n",
    "    return {\n",
    "            'data': data,\n",
    "            'layout': dict(\n",
    "            margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "            hovermode='x unified',\n",
    "            title={\n",
    "            'text': getTitleText(alcohol_title_text),\n",
    "            'y':0.9,\n",
    "            'x':0.5,\n",
    "            'xanchor': 'center',\n",
    "            'yanchor': 'top',\n",
    "            'font': dict(size=20),\n",
    "            },\n",
    "            hoverlabel=dict(\n",
    "                bgcolor=\"white\",\n",
    "                font=dict(\n",
    "                    size=16,\n",
    "                    family=\"Rockwell\"\n",
    "                ),\n",
    "            ),\n",
    "            legend=dict(\n",
    "                title=dict(\n",
    "                    text=\"Alcoholic beverages\",\n",
    "                )\n",
    "            ),\n",
    "        )\n",
    "    }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 455,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Alcohol sales callback\n",
    "@app.callback(Output('alcohol-stats', 'figure'), Input('alcohol-stats-type', 'value'))\n",
    "def update_alcohol_sales_plot(alcohol_stats_type):\n",
    "    if (alcohol_stats_type == datasets_options[0]):\n",
    "        return get_alcohol_sales_previous_year()\n",
    "    else:\n",
    "        return get_alcohol_sales_trend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 456,
   "metadata": {},
   "outputs": [],
   "source": [
    "trips_hovertemplate=\"%{y} million<extra></extra>\"\n",
    "trips_title_text = \"Domestic trips\"\n",
    "\n",
    "def get_trips_previous_year():\n",
    "    x=domestic_trips[\"type of destination\"]\n",
    "    q3_2019 = \"Q3 2019\"\n",
    "    q3_2020 = \"Q3 2020\"\n",
    "    \n",
    "    return {\n",
    "        'data': [dict(\n",
    "            x=x,\n",
    "            y=domestic_trips[\"Q3 2019\"],\n",
    "            type='bar',\n",
    "            marker={\n",
    "                'color': 'lime'\n",
    "            },\n",
    "            hovertemplate=q3_2019 +\"<br>\"+ trips_hovertemplate,\n",
    "            name=q3_2019,\n",
    "        ),\n",
    "        dict(\n",
    "            x=x,\n",
    "            y=domestic_trips[\"Q3 2020\"],\n",
    "            type='bar',\n",
    "            marker={\n",
    "                'color': 'green'\n",
    "            },\n",
    "            hovertemplate=q3_2020 +\"<br>\"+ trips_hovertemplate,\n",
    "            name=q3_2020,\n",
    "        )],\n",
    "        'layout': dict(\n",
    "            margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "            hovermode='closest',\n",
    "            title={\n",
    "            'text': getTitleText(trips_title_text),\n",
    "            'y':0.9,\n",
    "            'x':0.5,\n",
    "            'xanchor': 'center',\n",
    "            'yanchor': 'top',\n",
    "            'font': dict(size=20),\n",
    "            },\n",
    "            hoverlabel=dict(\n",
    "                bgcolor=\"white\",\n",
    "                font=dict(\n",
    "                    size=16,\n",
    "                    family=\"Rockwell\"\n",
    "                ),\n",
    "            ),\n",
    "            legend=dict(\n",
    "                    title=dict(\n",
    "                        text=\"Quarters\",\n",
    "                        )\n",
    "                    ),\n",
    "        )\n",
    "    }\n",
    "\n",
    "def get_trips_trend():\n",
    "    data = []\n",
    "    \n",
    "    for type_of_destination in trips_grouped.columns.get_level_values(\"type of destination\").unique():\n",
    "        trips = trips_grouped[type_of_destination][\"Total trips domestic\"]\n",
    "        sub_data = dict(\n",
    "            x=trips_grouped.index,\n",
    "            y=trips,\n",
    "            type='scatter',\n",
    "            name=type_of_destination,\n",
    "            hovertemplate=type_of_destination +\" : \"+ trips_hovertemplate,\n",
    "        )\n",
    "        data.append(sub_data)\n",
    "        \n",
    "    return {\n",
    "            'data': data,\n",
    "            'layout': dict(\n",
    "            margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "            hovermode='x unified',\n",
    "            title={\n",
    "            'text': getTitleText(trips_title_text),\n",
    "            'y':0.9,\n",
    "            'x':0.5,\n",
    "            'xanchor': 'center',\n",
    "            'yanchor': 'top',\n",
    "            'font': dict(size=20),\n",
    "            },\n",
    "            hoverlabel=dict(\n",
    "                bgcolor=\"white\",\n",
    "                font=dict(\n",
    "                    size=16,\n",
    "                    family=\"Rockwell\"\n",
    "                ),\n",
    "            ),\n",
    "            legend=dict(\n",
    "                title=dict(\n",
    "                    text=\"Type of destination\",\n",
    "                )\n",
    "            ),\n",
    "        )\n",
    "    }\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 457,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Trips callback\n",
    "@app.callback(Output('trips-stats', 'figure'), Input('trips-stats-type', 'value'))\n",
    "def update_alcohol_sales_plot(alcohol_stats_type):\n",
    "    if (alcohol_stats_type == datasets_options[0]):\n",
    "        return get_trips_previous_year()\n",
    "    else:\n",
    "        return get_trips_trend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 458,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Comparison tab callback\n",
    "@app.callback(Output('comparison-plot', 'figure'), [\n",
    "        Input('multi-country-input', 'value'),\n",
    "        Input('comparison-type', 'value'),\n",
    "    ])\n",
    "def update_comparison_plot(locations, comparison_type):\n",
    "    data = []\n",
    "    \n",
    "    for location in locations:\n",
    "        dfc = df[df.location == location]\n",
    "        sub_data = dict(\n",
    "            x=dfc.date,\n",
    "            y=dfc[comparison_type],\n",
    "            type='scatter',\n",
    "            name=location,\n",
    "            hovertemplate=\"%{y:,.2f}\",\n",
    "        )\n",
    "        data.append(sub_data)\n",
    "        \n",
    "    legend = dict(\n",
    "        title=dict(\n",
    "            text=\"Selected Locations\",\n",
    "        )\n",
    "    )\n",
    "        \n",
    "    return {\n",
    "            'data': data,\n",
    "            'layout': dict(\n",
    "            margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "            hovermode='x unified',\n",
    "            title={\n",
    "            'text': getTitleText(comparison_type) + \" (7-day smoothed)\",\n",
    "            'y':0.9,\n",
    "            'x':0.5,\n",
    "            'xanchor': 'center',\n",
    "            'yanchor': 'top',\n",
    "            'font': dict(size=20),\n",
    "            },\n",
    "            hoverlabel=dict(\n",
    "                bgcolor=\"white\",\n",
    "                font=dict(\n",
    "                    size=16,\n",
    "                    family=\"Rockwell\"\n",
    "                ),\n",
    "            ),\n",
    "                legend=legend,\n",
    "        )\n",
    "    }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 459,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Map callback\n",
    "@app.callback(\n",
    "    Output('world-map', 'figure'),\n",
    "    Input('map-type', 'value')\n",
    ")\n",
    "def display_map(map_type):\n",
    "    title = getTitleText(map_type)\n",
    "    \n",
    "    if(map_type == \"total_cases_per_million\"):\n",
    "        text = \"Global COVID-19 cases per million\"\n",
    "        column = dfm.total_cases_per_million\n",
    "    else:\n",
    "        text = \"Global COVID-19 deaths per million\"\n",
    "        column = dfm.total_deaths_per_million\n",
    "    \n",
    "    return {\n",
    "        'data': [dict(\n",
    "            locations= dfm.iso_code,\n",
    "            z = column.fillna(0),\n",
    "            text = dfm.location,\n",
    "            type='choropleth',\n",
    "            colorscale = 'Reds',\n",
    "            autocolorscale=False,\n",
    "            reversescale=False,\n",
    "            marker=dict(\n",
    "                line_color='darkgray',\n",
    "                line_width=0.5,\n",
    "            ),\n",
    "            colorbar = dict(\n",
    "                title = 'COVID-19 <br>'+title,\n",
    "            ),\n",
    "            geo=dict(\n",
    "                showframe=False,\n",
    "                showcoastlines=False,\n",
    "                projection_type='equirectangular'\n",
    "            ),\n",
    "            hoverinfo='z+text',\n",
    "            hovertemplate=\"%{z:,.2f}<br>%{text}<extra></extra>\"\n",
    "        )],\n",
    "        'layout': dict(\n",
    "            margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "            hovermode='closest',\n",
    "            title={\n",
    "            'text': text,\n",
    "            'y':0.9,\n",
    "            'x':0.5,\n",
    "            'xanchor': 'center',\n",
    "            'yanchor': 'top',\n",
    "            'font': dict(size=20)\n",
    "            },\n",
    "        )\n",
    "    }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 460,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_click_data(clickData):\n",
    "    int_formatter = \"{:,.0f}\".format\n",
    "    float_formatter = \"{:,.2f}\".format\n",
    "    location = None\n",
    "    dfcountry = None\n",
    "       \n",
    "    if(clickData != None): \n",
    "        data = clickData[\"points\"][0]\n",
    "        location = data[\"text\"]\n",
    "        country_code = data[\"location\"]\n",
    "        dfcountry = dfmain[dfmain.iso_code == country_code]\n",
    "    else:\n",
    "        location = \"World\"\n",
    "        dfcountry = dfmain[dfmain.iso_code == world_code]\n",
    "        \n",
    "    return (location, dfcountry.total_cases.apply(int_formatter), dfcountry.total_cases_per_million.apply(float_formatter), \n",
    "            dfcountry.total_deaths.apply(int_formatter), dfcountry.total_deaths_per_million.apply(float_formatter))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 461,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Map click callback\n",
    "@app.callback([\n",
    "    Output('click-data', 'children'),\n",
    "    Output('total-cases', 'children'),\n",
    "    Output('total-cases-pm', 'children'),\n",
    "    Output('total-deaths', 'children'),\n",
    "    Output('total-deaths-pm', 'children'),\n",
    "],\n",
    "    [\n",
    "        Input('world-map', 'clickData'),\n",
    "        Input('btn-world', 'n_clicks'),\n",
    "    ])\n",
    "def display_click_data(clickData, clicks):\n",
    "    changed_id = [p['prop_id'] for p in dash.callback_context.triggered][0]\n",
    "    if 'btn-world.n_clicks' == changed_id:\n",
    "        return get_click_data(None)\n",
    "    else:\n",
    "        return get_click_data(clickData)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 462,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predictions tab callback\n",
    "@app.callback([\n",
    "    Output('prediction-new-cases', 'figure'),\n",
    "    Output('prediction-new-deaths', 'figure'),\n",
    "], [\n",
    "    Input('prediction-location', 'value'),\n",
    "    Input('prediction-column', 'value'),\n",
    "])\n",
    "def update_prediction_plots(location, prediction_column):\n",
    "    \n",
    "    predictions_location = predictions[predictions.location == location]\n",
    "    actual = predictions_location[predictions_location.predictor.isna()]\n",
    "    arima = predictions_location[predictions_location.predictor == \"ARIMA\"]\n",
    "    prophet = predictions_location[predictions_location.predictor == \"Prophet\"]\n",
    "    \n",
    "    color = \"crimson\"\n",
    "    pridiction_color = \"indigo\"\n",
    "    prophet_upper_column = \"upper_cases\"\n",
    "    prophet_lower_column = \"lower_cases\"\n",
    "    limit_color=\"lavender\"\n",
    "    column_type = \"cases\"\n",
    "    if(prediction_column == \"new_cases_smoothed\"):\n",
    "        color = \"crimson\"\n",
    "        pridiction_color = \"indigo\"\n",
    "        prophet_upper_column = \"upper_cases\"\n",
    "        prophet_lower_column = \"lower_cases\"\n",
    "        limit_color=\"lavender\"\n",
    "        column_type = \"cases\"\n",
    "    elif(prediction_column == \"new_deaths_smoothed\"):\n",
    "        color = \"darkslateblue\"\n",
    "        pridiction_color = \"brown\"\n",
    "        prophet_upper_column = \"upper_deaths\"\n",
    "        prophet_lower_column = \"lower_deaths\"\n",
    "        limit_color=\"bisque\"\n",
    "        column_type = \"deaths\"\n",
    "        \n",
    "    legend = dict(title=dict(text=getTitleText(prediction_column) + \"<br>(7-day smoothed)\",))\n",
    "    \n",
    "    hover_template = \"%{y:,.2f}<extra></extra> \"+ column_type\n",
    "        \n",
    "    actual_graph_data = dict(\n",
    "                x=actual.date,\n",
    "                y=actual[prediction_column],\n",
    "                type='scatter',\n",
    "                name=\"Actual\",\n",
    "                hovertemplate=hover_template,\n",
    "                line = dict(color=color),\n",
    "            )\n",
    "    \n",
    "    figure_prophet = {\n",
    "        'data': [\n",
    "            dict(\n",
    "                 x=prophet.date,\n",
    "                 y=replace_negatives(prophet[prophet_upper_column]),\n",
    "                 type='scatter',\n",
    "                 name=\"Prophet\",\n",
    "                 hovertemplate=hover_template,\n",
    "                 line = dict(color=limit_color),\n",
    "                 showlegend=False,\n",
    "            ),\n",
    "            dict(\n",
    "                 x=prophet.date,\n",
    "                 y=replace_negatives(prophet[prophet_lower_column]),\n",
    "                 type='scatter',\n",
    "                 name=\"Prophet\",\n",
    "                 hovertemplate=hover_template,\n",
    "                 line = dict(color=limit_color),\n",
    "                 showlegend=False,\n",
    "                 fill='tonexty',\n",
    "                 fillcolor = limit_color,\n",
    "            ),\n",
    "            dict(\n",
    "                 x=prophet.date,\n",
    "                 y=replace_negatives(prophet[prediction_column]),\n",
    "                 type='scatter',\n",
    "                 name=\"Prophet\",\n",
    "                 hovertemplate=hover_template,\n",
    "                 line = dict(color=pridiction_color, dash='dash'),\n",
    "            ),\n",
    "            actual_graph_data,\n",
    "        ],\n",
    "            'layout': dict(\n",
    "            margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "            hovermode='closest',\n",
    "            title={\n",
    "            'text': \"Prophet Prediction\",\n",
    "            'y':0.9,\n",
    "            'x':0.5,\n",
    "            'xanchor': 'center',\n",
    "            'yanchor': 'top',\n",
    "            'font': dict(size=20),\n",
    "            },\n",
    "            hoverlabel=dict(\n",
    "                bgcolor=\"white\",\n",
    "                font=dict(\n",
    "                    size=16,\n",
    "                    family=\"Rockwell\"\n",
    "                ),\n",
    "            ),\n",
    "                legend=legend\n",
    "        )\n",
    "    }\n",
    "    \n",
    "    figure_arima = {\n",
    "        'data': [\n",
    "            actual_graph_data,\n",
    "            dict(\n",
    "                 x=arima.date,\n",
    "                 y=replace_negatives(arima[prediction_column]),\n",
    "                 type='scatter',\n",
    "                 name=\"ARIMA\",\n",
    "                 hovertemplate=hover_template,\n",
    "                 line = dict(color=pridiction_color, dash='dash')\n",
    "            ),\n",
    "        ],\n",
    "            'layout': dict(\n",
    "            margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "            hovermode='closest',\n",
    "            title={\n",
    "            'text': \"ARIMA Prediction\",\n",
    "            'y':0.9,\n",
    "            'x':0.5,\n",
    "            'xanchor': 'center',\n",
    "            'yanchor': 'top',\n",
    "            'font': dict(size=20),\n",
    "            },\n",
    "            hoverlabel=dict(\n",
    "                bgcolor=\"white\",\n",
    "                font=dict(\n",
    "                    size=16,\n",
    "                    family=\"Rockwell\"\n",
    "                ),\n",
    "            ),\n",
    "            legend=legend  \n",
    "        )\n",
    "    }\n",
    "    \n",
    "    return figure_arima, figure_prophet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 463,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_dot_graph(analysis_type, hightlight, column, title_text, hoverTemplate, color, highlight_color):\n",
    "    dfd = dfmain_no_world[dfmain_no_world.location != hightlight]\n",
    "    dfh = dfmain_no_world[dfmain_no_world.location == hightlight]\n",
    "    \n",
    "    return {\n",
    "            'data': [\n",
    "                dict(\n",
    "                    x=dfd[analysis_type],\n",
    "                    y=dfd[column],\n",
    "                    type='scatter',\n",
    "                    marker={\n",
    "                            'color': color,\n",
    "                        },\n",
    "                    name=dfd.location,\n",
    "                    hovertemplate= dfd.location +\"<extra></extra><br>\"+ hoverTemplate,\n",
    "                    mode=\"markers\",\n",
    "                    showlegend=False,\n",
    "                ),\n",
    "                dict(\n",
    "                    x=dfh[analysis_type],\n",
    "                    y=dfh[column],\n",
    "                    type='scatter',\n",
    "                    marker={\n",
    "                            'color': highlight_color,\n",
    "                            \"size\": 18,\n",
    "                        },\n",
    "                    name=dfh.location,\n",
    "                    hovertemplate= dfh.location +\"<extra></extra><br>\"+ hoverTemplate,\n",
    "                    mode=\"markers\",\n",
    "                    showlegend=False,\n",
    "                )\n",
    "            ],\n",
    "            'layout': dict(\n",
    "            margin={'l': 40, 'b': 30, 't': 60, 'r': 0},\n",
    "            hovermode='closest',\n",
    "            title={\n",
    "                'text': title_text,\n",
    "                'y':0.9,\n",
    "                'x':0.5,\n",
    "                'xanchor': 'center',\n",
    "                'yanchor': 'top',\n",
    "                'font': dict(size=20),\n",
    "            },\n",
    "            hoverlabel=dict(\n",
    "                bgcolor=\"white\",\n",
    "                font=dict(\n",
    "                    size=16,\n",
    "                    family=\"Rockwell\"\n",
    "                ),\n",
    "            ),\n",
    "                xaxis=dict(\n",
    "                    title=getTitleText(analysis_type)\n",
    "                ),\n",
    "                yaxis=dict(\n",
    "                    title=getTitleText(column)\n",
    "                ),\n",
    "        )\n",
    "    }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 464,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dot plot callbacks\n",
    "@app.callback(Output('dot-plot', 'figure'),[\n",
    "    Input('dot-metric', 'value'),\n",
    "    Input('hightlight-input', 'value'),\n",
    "])\n",
    "def update_total_dot_graph(analysis_type, hightlight):\n",
    "    column = \"total_cases_per_million\"\n",
    "    \n",
    "    hoverTemplate = getTitleText(analysis_type) +\": %{x:,.2f}<br>%{y:,.2f} cases per million people\"\n",
    "    \n",
    "    if(analysis_type == \"population\"):\n",
    "        hoverTemplate = getTitleText(analysis_type) +\": %{x:,.0f}<br>%{y:,.2f} cases per million people\"\n",
    "    \n",
    "    title_text = getTitleText(analysis_type) +\" Vs Total COVID-19 Cases Per Million\"\n",
    "    \n",
    "    return get_dot_graph(analysis_type, hightlight, column, title_text, hoverTemplate, \"burlywood\", \"darkred\")\n",
    "\n",
    "@app.callback(Output('total-dot-plot', 'figure'),[\n",
    "    Input('dot-metric', 'value'),\n",
    "    Input('hightlight-input', 'value'),\n",
    "])\n",
    "def update_dot_graph(analysis_type, hightlight):\n",
    "    column = \"total_deaths_per_million\"\n",
    "    \n",
    "    hoverTemplate = getTitleText(analysis_type) +\": %{x:,.2f}<br>%{y:,.2f} deaths per million people\"\n",
    "    \n",
    "    if(analysis_type == \"population\"):\n",
    "        hoverTemplate = getTitleText(analysis_type) +\": %{x:,.0f}<br>%{y:,.2f} deaths per million people\"\n",
    "    \n",
    "    title_text = getTitleText(analysis_type) +\" Vs Total COVID-19 Deaths Per Million\"\n",
    "    \n",
    "    return get_dot_graph(analysis_type, hightlight, column, title_text, hoverTemplate, \"deepskyblue\", \"darkviolet\")\n",
    "\n",
    "@app.callback([\n",
    "    Output('hightlight-input', 'options'),\n",
    "    Output('hightlight-input', 'value'),\n",
    "], Input('dot-metric', 'value'))\n",
    "def update_highlight_locations(analysis_type):\n",
    "    analysis_locations = dfmain_no_world[~dfmain_no_world[analysis_type].isna()].location\n",
    "    analysis_locations_list = analysis_locations.tolist()\n",
    "    first_location = \"\" if len(analysis_locations_list) == 0 else analysis_locations_list[0]\n",
    "    default_location = \"Norway\" if \"Norway\" in analysis_locations_list else first_location\n",
    "    \n",
    "    return [\n",
    "        {'label': location, 'value': location} for location in analysis_locations\n",
    "    ], default_location"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 465,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reference: https://github.com/nachi-hebbar/ARIMA-Temperature_Forecasting\n",
    "\n",
    "def predict_arima(df, country, column):\n",
    "    from statsmodels.tsa.arima_model import ARIMA\n",
    "    from pmdarima import auto_arima\n",
    "    import warnings\n",
    "    df=df[df.location == country]\n",
    "    df = df[[column]]\n",
    "    df=df.dropna()\n",
    "    print('Shape of data',df.shape)\n",
    "\n",
    "    #Figure Out Order for ARIMA Model\n",
    "    # Ignore harmless warnings\n",
    "    warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "    stepwise_fit = auto_arima(df[column], \n",
    "                              suppress_warnings=True)           \n",
    "\n",
    "    order = stepwise_fit.order\n",
    "\n",
    "    model2=ARIMA(df[column],order=order)\n",
    "    model2=model2.fit()\n",
    "    \n",
    "    predict_start_date = df[-1:].index[0] + pd.DateOffset(1)\n",
    "\n",
    "    index_future_dates=pd.date_range(predict_start_date, periods=180)\n",
    "\n",
    "    prediction=model2.predict(start=len(df),end=len(df) + index_future_dates.size - 1 ,\n",
    "                              typ='levels').rename('ARIMA Predictions')\n",
    "    prediction.index=index_future_dates\n",
    "    return prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 466,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reference: https://machinelearningmastery.com/time-series-forecasting-with-prophet-in-python/\n",
    "\n",
    "def predict_prophet(df, country, column):\n",
    "    from fbprophet import Prophet    \n",
    "    df=df[df.location == country]\n",
    "    df = df[[\"date\", column]]\n",
    "    df = df.rename(columns={\"date\":\"ds\", column:\"y\"})\n",
    "    df=df.dropna()\n",
    "    print('Shape of data',df.shape)\n",
    "    # define the model\n",
    "    model = Prophet()\n",
    "    # fit the model\n",
    "    model.fit(df)\n",
    "\n",
    "    predict_start_date = df.iloc[-1].ds + pd.DateOffset(1)\n",
    "\n",
    "    future=pd.date_range(predict_start_date, periods=180)\n",
    "\n",
    "    future = pd.DataFrame(future)\n",
    "    future.columns = ['ds']\n",
    "    # use the model to make a forecast\n",
    "    forecast = model.predict(future)\n",
    "    return forecast[\"yhat\"], forecast[\"yhat_upper\"], forecast[\"yhat_lower\"],"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 467,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict():    \n",
    "    dfd = df.set_index(\"date\")\n",
    "    dfd.index = pd.to_datetime(dfd.index)\n",
    "    \n",
    "    columns = [\"new_cases_smoothed\", \"new_deaths_smoothed\"]\n",
    "    \n",
    "    countries = [\"Norway\", \"Sweden\", \"Finland\", \"Denmark\", \"Iceland\"]\n",
    "    dfd = dfd[dfd.location.isin(countries)]\n",
    "    \n",
    "    pred_max = pd.to_datetime(predictions[predictions.predictor.isna()].date.max())\n",
    "    df_max = pd.to_datetime(dfd.index.max())\n",
    "    if(pred_max == df_max):\n",
    "        return \"Up to date\"\n",
    "\n",
    "    dfr = dfd.reset_index()\n",
    "    dfp = dfr[[\"date\", \"location\", \"new_cases_smoothed\", \"new_deaths_smoothed\"]].copy()\n",
    "    dfp[\"predictor\"] = \"\"\n",
    "    dfp[\"upper_deaths\"] = \"\"\n",
    "    dfp[\"lower_deaths\"] = \"\"\n",
    "    dfp[\"upper_cases\"] = \"\"\n",
    "    dfp[\"lower_cases\"] = \"\"\n",
    "    \n",
    "    for country in countries:\n",
    "        pred1 = predict_arima(dfd, country, \"new_cases_smoothed\")\n",
    "        pred2 = predict_arima(dfd, country, \"new_deaths_smoothed\")\n",
    "        dfi = pd.DataFrame({ \"new_cases_smoothed\": pred1, \"new_deaths_smoothed\": pred2})\n",
    "        dfi[\"date\"] = pred1.index\n",
    "        dfi[\"location\"] = country\n",
    "        dfi[\"predictor\"] = \"ARIMA\"\n",
    "        dfp = dfp.append(dfi, ignore_index=True)\n",
    "\n",
    "        pred3, preduc, predlc = predict_prophet(dfr, country, \"new_cases_smoothed\")\n",
    "        pred4, predud, predld = predict_prophet(dfr, country, \"new_deaths_smoothed\")\n",
    "        dfj = pd.DataFrame({ \"new_cases_smoothed\": pred3, \"new_deaths_smoothed\": pred4,\n",
    "                           \"upper_deaths\": predud, \"lower_deaths\": predld, \"upper_cases\": preduc, \"lower_cases\": predlc})\n",
    "        dfj[\"date\"] = pred1.index\n",
    "        dfj[\"location\"] = country\n",
    "        dfj[\"predictor\"] = \"Prophet\"\n",
    "        dfp = dfp.append(dfj, ignore_index=True)\n",
    "        \n",
    "    dfp.to_csv(\"predictions.csv\", index=False)\n",
    "    return \"Updated\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 468,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dash app running on http://127.0.0.1:8050/\n"
     ]
    }
   ],
   "source": [
    "app.run_server(debug=debug)\n",
    "# app.run_server(debug=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 469,
   "metadata": {},
   "outputs": [],
   "source": [
    "# predict()"
   ]
  }
 ],
 "metadata": {
  "jupytext": {
   "formats": "ipynb,py:percent"
  },
  "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}