0

I am currently using the wetterdienst module with the DWD (Deutscher Wetterdienst, Germany's National Meteorological Service) server for retrieving e.g. MOSMIX data (weather forecasts).

For training an ML model with forecasts dating back more than a few days I need to save them locally as they are not archived on the DWD Open Data server.

But how can I read them into the wetterdienst module again? The API only provides for direct loading of files from the server (with local dogpile caching). I want to use the existing parts of wetterdienst for processing the KMZ files I saved earlier.

I cannot use actual historic observation data as they lack certain forecast data fields, e.g. precipitation probabilities.

Stefan
  • 1,036
  • 10
  • 32
  • 1
    The question doesn't really make sense to me. Once you have a `pandas.DataFrame` from `request.values.all().df`, all of the work is done in pandas. The wetterdienst API seems to just be for downloading the data. If you have data that has been saved locally from pandas. then just reload it into pandas with one of the `pd.read_` methods. – Trenton McKinney Jun 14 '21 at 17:33
  • @TrentonMcKinney It's a question of platform availability. I'd have to prepare a proper Python & Co. environment on the target system used for archiving the files. Then your suggestion is a good alternative to storing the original KMZ files issued by DWD. – Stefan Jun 15 '21 at 07:40

0 Answers0