francois/HowBadIsMyBatch
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactoring

Browse Source
This commit is contained in:
frankknoll
2022-02-05 17:39:46 +01:00
parent a8c20c3656
commit 9f9ae08d3c
1 changed files with 8 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
HowBadIsMyBatch.ipynb
View File

@@ -261,35 +261,23 @@
" \n",
" @staticmethod\n",
" def getDoseTable(dataFrame):\n",
" doseTable = AggregationHelper.aggregateAndFlattenColumnsAndRenameColumns(\n",
" dataFrame = dataFrame.groupby('VAX_DOSE_SERIES'),\n",
" aggFunctionsByColumn = {\n",
" 'DIED': ['sum', 'size'],\n",
" 'L_THREAT': 'sum',\n",
" 'DISABLE': 'sum'\n",
" },\n",
" columnNameMappingsDict = {\n",
" \"DIED_size\": \"Total reports\",\n",
" \"DIED_sum\": \"Deaths\",\n",
" \"L_THREAT_sum\": \"Life Threatening Illnesses\",\n",
" \"DISABLE_sum\": \"Disabilities\"\n",
" })\n",
" doseTable = doseTable[['Total reports', 'Deaths', 'Disabilities', 'Life Threatening Illnesses']]\n",
" doseTable['Severe reports (%)'] = (doseTable['Deaths'] + doseTable['Disabilities'] + doseTable['Life Threatening Illnesses']) / doseTable['Total reports'] * 100\n",
" return doseTable\n",
" return DoseAnalysis._getDoseTable(dataFrame.groupby('VAX_DOSE_SERIES'))\n",
"\n",
" @staticmethod\n",
" def getDoseByMonthTable(dataFrame):\n",
" # https://stackoverflow.com/questions/61879166/pandas-groupby-month-and-year-date-as-datetime64ns-and-summarized-by-count\n",
" grouped = dataFrame.groupby(\n",
" return DoseAnalysis._getDoseTable(\n",
" dataFrame.groupby(\n",
" [\n",
" dataFrame['RECVDATE'].dt.year.rename('year'),\n",
" dataFrame['RECVDATE'].dt.month.rename('month'),\n",
" dataFrame['VAX_DOSE_SERIES']\n",
" ])\n",
" # FK-TODO: DRY with getDoseTable()\n",
" ]))\n",
"\n",
" @staticmethod\n",
" def _getDoseTable(dataFrame):\n",
" doseTable = AggregationHelper.aggregateAndFlattenColumnsAndRenameColumns(\n",
" dataFrame = grouped,\n",
" dataFrame = dataFrame,\n",
" aggFunctionsByColumn = {\n",
" 'DIED': ['sum', 'size'],\n",
" 'L_THREAT': 'sum',\n",