adding Einwohnerzahl

src/help.txt

@@ -8,6 +8,8 @@ zwei Spalten darstellen:
 
 src/intensivstationen/AGS_2022-02-28.json downloaded from https://www.xrepository.de/details/urn:de:bund:destatis:bevoelkerungsstatistik:schluessel:ags
 
+04-kreise.xlsx: https://www.destatis.de/DE/Themen/Laender-Regionen/Regionales/Gemeindeverzeichnis/Administrativ/04-kreise.html
+
 get VAERS data:
 - download data (e.g. 2022VAERSData.zip) from https://vaers.hhs.gov/data/datasets.html and save and unzip in VAERS folder
 

src/intensivstationen/Intensivstationen.ipynb

@@ -17,29 +17,22 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "f7868050",
+   "id": "0927a6c2",
    "metadata": {},
    "outputs": [],
    "source": [
-    "def readKreisByKreisschluessel():\n",
-    "    kreisByKreisschluessel = pd.read_csv(\n",
-    "        'kreisByKreisschluessel.csv',\n",
-    "        low_memory = False,\n",
-    "        index_col = 'Kreisschluessel',\n",
-    "        dtype = {'Kreisschluessel': 'string'})\n",
-    "    kreisByKreisschluessel['Kreis'] = kreisByKreisschluessel['Kreis'].str.strip()\n",
-    "    return kreisByKreisschluessel[kreisByKreisschluessel.index.str.len() == 5]\n"
+    "kreise = pd.read_excel('04-kreise.xlsx', sheet_name = 1, header = 5, index_col = 0)\n",
+    "kreise"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "712a919c",
+   "id": "3f75edff",
    "metadata": {},
    "outputs": [],
    "source": [
-    "kreisByKreisschluessel = readKreisByKreisschluessel()\n",
-    "kreisByKreisschluessel"
+    "kreise.columns"
    ]
   },
   {
@@ -83,31 +76,21 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "class KreisAndBundeslandColumnAdder:\n",
+    "class ColumnsAdder:\n",
     "\n",
-    "    def __init__(self, kreisByKreisschluessel):\n",
-    "        self.kreisByKreisschluessel = kreisByKreisschluessel\n",
+    "    def __init__(self, kreise):\n",
+    "        self.kreise = kreise\n",
     "\n",
-    "    def addKreisAndBundeslandColumn(self, dataFrame):\n",
+    "    def addKreisAndBundeslandAndEinwohnerzahlColumns(self, dataFrame):\n",
     "        dataFrame['Kreis'] = dataFrame['gemeindeschluessel'].map(\n",
-    "            lambda gemeindeschluessel: self.kreisByKreisschluessel.loc[gemeindeschluessel, 'Kreis'])\n",
-    "        dataFrame['Bundesland'] = dataFrame['bundesland'].map({\n",
-    "            '01': 'Schleswig-Holstein',\n",
-    "            '02': 'Hamburg',\n",
-    "            '03': 'Niedersachsen',\n",
-    "            '04': 'Bremen',\n",
-    "            '05': 'Nordrhein-Westfalen',\n",
-    "            '06': 'Hessen',\n",
-    "            '07': 'Rheinland-Pfalz',\n",
-    "            '08': 'Baden-Württemberg',\n",
-    "            '09': 'Bayern',\n",
-    "            '10': 'Saarland',\n",
-    "            '11': 'Berlin',\n",
-    "            '12': 'Brandenburg',\n",
-    "            '13': 'Mecklenburg-Vorpommern',\n",
-    "            '14': 'Sachsen',\n",
-    "            '15': 'Sachsen-Anhalt',\n",
-    "            '16': 'Thüringen'})\n",
+    "            lambda gemeindeschluessel: self.kreise.loc[gemeindeschluessel, 3])\n",
+    "\n",
+    "        dataFrame['Bundesland'] = dataFrame['bundesland'].map(\n",
+    "            lambda bundesland: self.kreise.loc[bundesland, '2'])\n",
+    "\n",
+    "        dataFrame['Einwohnerzahl'] = dataFrame['gemeindeschluessel'].map(\n",
+    "            lambda gemeindeschluessel: int(self.kreise.loc[gemeindeschluessel, 6]))\n",
+    "\n",
     "        return dataFrame\n"
    ]
   },
@@ -118,7 +101,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "timeSeries = KreisAndBundeslandColumnAdder(kreisByKreisschluessel).addKreisAndBundeslandColumn(timeSeries)\n",
+    "timeSeries = ColumnsAdder(kreise).addKreisAndBundeslandAndEinwohnerzahlColumns(timeSeries)\n",
     "timeSeries"
    ]
   },