{ "cells": [ { "cell_type": "code", "execution_count": null, "id": "9de5907f-18f5-4cb1-903e-26028ff1fa03", "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "import pandas as pd\n", "from urllib import request\n", "\n", "pd.set_option('display.max_rows', 100)\n", "pd.set_option('display.max_columns', None)" ] }, { "cell_type": "code", "execution_count": null, "id": "d021de84", "metadata": {}, "outputs": [], "source": [ "# download https://diviexchange.blob.core.windows.net/%24web/zeitreihe-tagesdaten.csv or https://www.intensivregister.de/#/aktuelle-lage/downloads\n", "if False:\n", " request.urlretrieve(\n", " 'https://diviexchange.blob.core.windows.net/%24web/zeitreihe-tagesdaten.csv',\n", " 'zeitreihe-tagesdaten.csv')\n" ] }, { "cell_type": "code", "execution_count": null, "id": "f179762b", "metadata": {}, "outputs": [], "source": [ "def readTimeseries():\n", " timeseries = pd.read_csv(\n", " 'zeitreihe-tagesdaten.csv',\n", " low_memory = False,\n", " usecols = ['date', 'bundesland', 'gemeindeschluessel', 'betten_belegt', 'betten_frei'],\n", " parse_dates = ['date'],\n", " date_parser = lambda dateStr: pd.to_datetime(dateStr, format = \"%Y-%m-%d\"),\n", " dtype = {\n", " 'gemeindeschluessel': 'string',\n", " 'bundesland': 'string'\n", " })\n", " return timeseries.sort_values(by = 'date', ascending = True)\n" ] }, { "cell_type": "code", "execution_count": null, "id": "2d34c6a4", "metadata": {}, "outputs": [], "source": [ "# download https://www.destatis.de/DE/Themen/Laender-Regionen/Regionales/Gemeindeverzeichnis/Administrativ/04-kreise.xlsx?__blob=publicationFile or https://www.destatis.de/DE/Themen/Laender-Regionen/Regionales/Gemeindeverzeichnis/Administrativ/04-kreise.html\n", "if False:\n", " request.urlretrieve(\n", " 'https://www.destatis.de/DE/Themen/Laender-Regionen/Regionales/Gemeindeverzeichnis/Administrativ/04-kreise.xlsx?__blob=publicationFile',\n", " '04-kreise.xlsx')" ] }, { "cell_type": "code", "execution_count": null, "id": "0927a6c2", "metadata": {}, "outputs": [], "source": [ "def readKreise():\n", " kreise = pd.read_excel(\n", " '04-kreise.xlsx',\n", " sheet_name = 'Kreisfreie Städte u. Landkreise',\n", " header = 5,\n", " index_col = 0)\n", " kreise = kreise.rename(columns = {'2': 'Bundesland', 3: 'Kreis', 6: 'Einwohnerzahl'})[['Bundesland', 'Kreis', 'Einwohnerzahl']]\n", " kreise.index.set_names(\"Key\", inplace = True)\n", " return kreise" ] }, { "cell_type": "code", "execution_count": null, "id": "af96fb11", "metadata": {}, "outputs": [], "source": [ "class ColumnsAdder:\n", "\n", " def __init__(self, kreise):\n", " self.kreise = kreise\n", "\n", " def addKreisAndBundeslandAndEinwohnerzahlColumns(self, dataFrame):\n", " dataFrame_kreise = pd.merge(dataFrame, self.kreise, how = 'left', left_on = 'gemeindeschluessel', right_index = True)\n", " dataFrame['Kreis'] = dataFrame_kreise['Kreis']\n", " dataFrame['Einwohnerzahl'] = dataFrame_kreise['Einwohnerzahl']\n", " return self._addBundeslandColumn(dataFrame)\n", " \n", " def _addBundeslandColumn(self, dataFrame):\n", " return pd.merge(\n", " dataFrame,\n", " self._createBundeslandByKeyTable(),\n", " how = 'left',\n", " left_on = 'bundesland',\n", " right_index = True)\n", "\n", " def _createBundeslandByKeyTable(self):\n", " return self.kreise[self.kreise.index.str.len() == 2][['Bundesland']]\n" ] }, { "cell_type": "code", "execution_count": null, "id": "62a20115", "metadata": {}, "outputs": [], "source": [ "timeSeries = ColumnsAdder(readKreise()).addKreisAndBundeslandAndEinwohnerzahlColumns(readTimeseries())\n", "timeSeries" ] }, { "cell_type": "code", "execution_count": null, "id": "356494d3", "metadata": {}, "outputs": [], "source": [ "kreisValues = timeSeries['Kreis'].drop_duplicates().values\n", "kreisValues" ] }, { "cell_type": "code", "execution_count": null, "id": "05aa0117", "metadata": {}, "outputs": [], "source": [ "def printKreisOptions(kreisValues):\n", " for kreis in kreisValues:\n", " printKreisOption(kreis)\n", "\n", "def printKreisOption(kreis):\n", " print(''.format(kreis = kreis))" ] }, { "cell_type": "code", "execution_count": null, "id": "33a4b725", "metadata": {}, "outputs": [], "source": [ "kreisValues = sorted(kreisValues)\n", "printKreisOptions(kreisValues)" ] }, { "cell_type": "code", "execution_count": null, "id": "43c2f826", "metadata": {}, "outputs": [], "source": [ "import os\n", "\n", "\n", "class IOUtils:\n", "\n", " def saveDictAsJson(dict, file):\n", " IOUtils.ensurePath(file)\n", " with open(file, 'w') as outfile:\n", " json.dump(dict, outfile)\n", "\n", " @staticmethod\n", " def ensurePath(file):\n", " directory = os.path.dirname(file)\n", " if not os.path.exists(directory):\n", " os.makedirs(directory)\n" ] }, { "cell_type": "code", "execution_count": null, "id": "997a4bdb", "metadata": {}, "outputs": [], "source": [ "def getIntensiveCareBeds(timeSeries, kreis = None):\n", " if kreis is not None:\n", " return timeSeries[timeSeries['Kreis'] == kreis][['date', 'betten_belegt', 'betten_frei', 'Einwohnerzahl']]\n", " else:\n", " return timeSeries.groupby('date').agg(**{\n", " 'betten_belegt': pd.NamedAgg(column = 'betten_belegt', aggfunc = 'sum'),\n", " 'betten_frei': pd.NamedAgg(column = 'betten_frei', aggfunc = 'sum'),\n", " 'Einwohnerzahl': pd.NamedAgg(column = 'Einwohnerzahl', aggfunc = 'sum') \n", " }).reset_index()" ] }, { "cell_type": "code", "execution_count": null, "id": "a97f5b2b", "metadata": {}, "outputs": [], "source": [ "import json\n", "\n", "\n", "def getAndPersistIntensiveCareBeds(timeSeries, kreis=None):\n", " dataFrame = getIntensiveCareBeds(timeSeries, kreis)\n", " display(kreis)\n", " _saveDataFrameAsJson(dataFrame, _getFilename(kreis))\n", " return dataFrame\n", "\n", "\n", "def _saveDataFrameAsJson(dataFrame, file):\n", " IOUtils.saveDictAsJson(\n", " {\n", " 'population': int(dataFrame.iloc[0]['Einwohnerzahl']),\n", " 'data': _dataFrame2Dict(dataFrame),\n", " },\n", " file)\n", "\n", "\n", "def _dataFrame2Dict(dataFrame):\n", " df = dataFrame[['date', 'betten_belegt', 'betten_frei']]\n", " df['date'] = df['date'].dt.strftime('%Y-%m-%d')\n", " return df.to_dict(orient=\"records\")\n", "\n", "\n", "def _getFilename(kreis):\n", " return '../../docs/data/intensivstationen/intensivstationen-{suffix}.json'.format(suffix=_getSuffix(kreis))\n", "\n", "\n", "def _getSuffix(kreis):\n", " return kreis if kreis is not None else 'de'\n" ] }, { "cell_type": "code", "execution_count": null, "id": "349edd73", "metadata": {}, "outputs": [], "source": [ "getAndPersistIntensiveCareBeds(timeSeries)" ] }, { "cell_type": "code", "execution_count": null, "id": "1b97137f", "metadata": {}, "outputs": [], "source": [ "for kreis in kreisValues:\n", " getAndPersistIntensiveCareBeds(timeSeries, kreis)" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "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": "3.9.7" } }, "nbformat": 4, "nbformat_minor": 5 }