Animated Map - 7-Day Rolling Average of Positive Tests.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Already up to date.\r\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "%config InlineBackend.print_figure_kwargs={'facecolor' : \"#1f1e1e\"}\n",
    "\n",
    "import io\n",
    "import os\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import random\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import plotly.express as px\n",
    "import plotly.graph_objects as go\n",
    "from datetime import datetime, timedelta\n",
    "from IPython.display import Markdown, display, Image\n",
    "from PIL import Image as PILImage\n",
    "\n",
    "# LOCAL GIT REPOSITORY LOCATION FOR https://github.com/CSSEGISandData/COVID-19\n",
    "local_covid_19_data = '/Users/davidmorton/Documents/COVID-19/'\n",
    "\n",
    "!git -C $local_covid_19_data pull\n",
    "\n",
    "states_to_drop = ['Puerto Rico', 'American Samoa', 'Guam', 'District of Columbia', 'Northern Mariana Islands', 'Virgin Islands']\n",
    "\n",
    "state_names = pd.read_csv('Data/states.csv').set_index('State').to_dict()['Abbreviation']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_single_date_data(date):\n",
    "    datestr = date.strftime('%m-%d-%Y')\n",
    "    filename = local_covid_19_data + 'csse_covid_19_data/csse_covid_19_daily_reports_us/' + datestr + '.csv'\n",
    "    result = pd.read_csv(filename)\n",
    "    result['Date'] = date\n",
    "    return result\n",
    "\n",
    "def get_daily_report_data():\n",
    "    start_date_for_daily_data = datetime(2020, 4, 12)\n",
    "    current_date = start_date_for_daily_data\n",
    "\n",
    "    yesterday = (datetime.now() + timedelta(days = -1))\n",
    "    #yesterday = datetime.now()\n",
    "    lastdate = yesterday.strftime('%m-%d-%Y')\n",
    "    \n",
    "    series = pd.DataFrame()\n",
    "\n",
    "    while True:\n",
    "        series = series.append(get_single_date_data(current_date))\n",
    "        current_date = current_date + timedelta(days = 1)\n",
    "        if current_date > yesterday:\n",
    "            break\n",
    "    \n",
    "    \n",
    "    return series\n",
    "\n",
    "def get_heatmap_data():\n",
    "    states_to_drop = ['Puerto Rico', 'American Samoa', 'Guam', 'Northern Mariana Islands', 'Virgin Islands', 'Diamond Princess', 'Grand Princess', 'Recovered']\n",
    "\n",
    "    # Get the daily report, and immediately sort it by province and date, and remove the non-state territories. \n",
    "    data = get_daily_report_data()[['Province_State', 'Date', 'Confirmed', 'People_Tested']]\n",
    "    data = data.sort_values(by=['Province_State', 'Date']).reset_index()\n",
    "    data = data[~data['Province_State'].isin(states_to_drop)]\n",
    "    data = data.set_index(['Province_State', 'Date'])\n",
    "    \n",
    "    # Create a confirmed percentage column, and then roll it in a 7-day rolling average\n",
    "    data['ConfirmedPercent'] = data['Confirmed'].div(data['People_Tested'])\n",
    "    data['ConfirmedPercentRolling'] = data.sort_values(by=['Province_State', 'Date']).groupby('Province_State')['ConfirmedPercent'].rolling(window=7).mean().reset_index(level=1, drop=1)\n",
    "\n",
    "    # Create a pivot table for use in the heatmap.\n",
    "    data = data.reset_index()[['Province_State', 'Date', 'ConfirmedPercentRolling']].pivot(values='ConfirmedPercentRolling', index='Date', columns='Province_State')\n",
    "    data = data.reindex(data.mean().sort_values(ascending=False).index, axis=1).dropna()\n",
    "\n",
    "    return data\n",
    "\n",
    "def draw_figure():\n",
    "    xxp = get_heatmap_data()\n",
    "    \n",
    "    ylabels = [' {0} ({1:.2f}%) '.format(a[0], a[1]*100) for a in zip(xxp.columns, xxp.mean().sort_values(ascending=False))]\n",
    "    xlabels = [list(reversed(pd.Series(xxp.index)))[i].strftime('%B %-d') if i % 7==0 else '' for i in np.arange(len(xxp.index)-1,-1, -1)]\n",
    "\n",
    "    plt.style.use('dark_background')\n",
    "    plt.subplots(figsize=(18,23))\n",
    "\n",
    "    ax = sns.heatmap(xxp.transpose() * 100, \n",
    "                     vmin=0, \n",
    "                     vmax=20, \n",
    "                     cmap='icefire')\n",
    "\n",
    "    ax.set_xticklabels(labels=xlabels, rotation=45, ha='right', size=14)\n",
    "    ax.set_xlabel('')\n",
    "    ax.set_yticklabels(labels=ylabels, size=14)\n",
    "    ax.set_ylabel('')\n",
    "    ax.tick_params(axis='both', which='both', length=0)\n",
    "\n",
    "    ax.collections[0].colorbar.ax.tick_params(labelsize=14)\n",
    "\n",
    "    ax.set_title('7-Day Rolling Average of Positive COVID-19\\nTest Results Sorted by State Average', size=36, pad=20, fontname='Helvetica')\n",
    "    \n",
    "    plt.annotate('Source: Johns Hopkins COVID-19 Data (https://github.com/CSSEGISandData/COVID-19)', (0,0), (100, -90), xycoords='axes fraction', fontname='Helvetica', textcoords='offset points', size=14, va='top')\n",
    "    plt.savefig(\"covid19-confirmed-heatmap.png\", bbox_inches=\"tight\", pad_inches=0.5)\n",
    "    \n",
    "def make_usa_choropleth(series, locations, title, colorbartitle, colorspectrum, zmax):\n",
    "    \n",
    "\n",
    "    fig = go.Figure(\n",
    "        data=go.Choropleth(\n",
    "            locations=locations,\n",
    "            z = series,\n",
    "            locationmode = 'USA-states',\n",
    "            colorscale = colorspectrum,\n",
    "            colorbar_title = colorbartitle, \n",
    "            zmax=zmax,\n",
    "            zmin=0\n",
    "        )\n",
    "    )\n",
    "\n",
    "    fig.update_layout(\n",
    "        title_text = title,\n",
    "        geo_scope='usa', # limite map scope to USA\n",
    "    )\n",
    "\n",
    "    return fig;\n",
    "\n",
    "def make_animation(totalduration, pausetime, timeinterpolation, theme):\n",
    "    bgcolor=\"black\"\n",
    "    data = get_heatmap_data() * 100\n",
    "    data = data.rename(state_names, axis=1)\n",
    "    \n",
    "    vals = data.values.flatten().flatten()\n",
    "    vals.sort()\n",
    "    bins = pd.qcut(vals, q=8, retbins=True)[1]\n",
    "    colorstops = (bins - bins.min()) / (bins.max() - bins.min())\n",
    "    colorscale = list(zip(colorstops, getattr(px.colors.sequential, theme)))\n",
    "    zmax = vals.max()\n",
    "\n",
    "    data = data.resample(rule=timeinterpolation).asfreq().interpolate()\n",
    "    \n",
    "    adjduration = totalduration-pausetime;\n",
    "    perslide = round((adjduration*1000)/len(data));\n",
    "    repeatlast = round((pausetime*1000)/perslide)\n",
    "    print('Frames will be shown for {0} milliseconds'.format(perslide))\n",
    "    \n",
    "    images = []\n",
    "    for i in range(len(data)):\n",
    "        row = data.iloc[i]\n",
    "        fig = make_usa_choropleth(row, row.index, 'Positive Percentage ' + row.name.strftime('%B %d') + ' ({:.2f}% Average)'.format(row.mean()), '%', colorscale, zmax)\n",
    "        fig.update_layout(\n",
    "            plot_bgcolor=bgcolor,\n",
    "            paper_bgcolor=bgcolor,\n",
    "            font=dict(color=\"white\"),\n",
    "            geo = dict(\n",
    "                lakecolor=bgcolor,\n",
    "                showlakes=True,\n",
    "                bgcolor=bgcolor\n",
    "            )\n",
    "        )\n",
    "        img_bytes = fig.to_image(format='png')\n",
    "        image = PILImage.open(io.BytesIO(img_bytes))\n",
    "\n",
    "        images.append(image)\n",
    "        if (i == (len(data) - 1)):\n",
    "            for j in range(1,repeatlast):\n",
    "                images.append(image)\n",
    "        \n",
    "    url = 'Output\\covid19-case-lapse.gif'\n",
    "    images[0].save(url,\n",
    "                   save_all=True, append_images=images[1:], dpi=300, optimize=True, duration=perslide, loop=0)\n",
    "    return url + \"?c=\" + str(random.randint(0,2e9))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Frames will be shown for 52 milliseconds\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<img src=\"covid19-case-lapse.gif?c=382954299\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "url = make_animation(30, 5, '6H', \"Inferno\")\n",
    "Image(url=url)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}