Scott Chamberlin

If opening with xarray open_mfdataset and parallel=True it will fail unless you have previously opened it with parallel=False

23

Minimal repro: ``` import xarray as xr ds = xr.open_mfdataset('gfs.0p25.201511*00.f0*.grib2', engine='cfgrib', combine='nested', concat_dim=['step'], parallel=True, chunks=24, backend_kwargs={'filter_by_keys': {'typeOfLevel': 'surface'}, 'indexpath': ''}) ``` Expected result: returns xarray Actual result: ``` ECCODES ERROR...

The current 5k training set is too small Double it to 10k and find the features from tsfresh with the highest importance to then map back in to the tsai...

Currently .25 lat/lon grid is generating too many similar elements per label. Explore: 1. Keep the .25 sample but only pick one element per region per date (try random selection...

Since we are using a regional forecast to train a point forecast it would be useful to have a view which allowed us to analyze the deviation within a forecast...

Currently there isn't a constent and long term accuracy tracking capability. Integrate one.