Aaron Spring

Multi-dimensional LogisticRegression

7

First of all: great software combining sklearn and xarray @phausamann My example here is inspired from #52, an example which helped me a lot. Input data: forecast with dimension T...

Some xskillscore metrics changed order of dims making CI doctests fail

Add bounds?

1

https://confluence.ecmwf.int/display/COPSRV/Guide+to+NetCDF+encoding+for+C3S+providers lead_bnds, valid_time_bnds, time_bnds For now user decides whether timestamp represents beginning/center/end of period

http://library.ucsd.edu/dc/object/bb88449405 Download 10GB zip file first. ```python z = xr.open_dataset("MJO_ECMWFzar",engine='zarr') mi = pd.MultiIndex.from_arrays([z.Ense.values, z.Forecast_Date.values, z.Forecast_Lead.values],names=("member","init","lead")) ds = z.assign_coords(mi=("index",mi)).swap_dims({"index":"mi"}).drop("index").unstack().drop(["Ense","Forecast_Date","Forecast_Lead","Center"]) # convert init to cftime/datetime ```

allow `datetime` after `__init__`

2

# Description - resolves logging error when working with `datetime` internally after: ```python # convert to datetime hindcast._datasets['initialized'] = hindcast._datasets['initialized'].convert_calendar('standard',dim='init') hindcast._datasets['initialized']['valid_time'] = xr.concat([hindcast._datasets['initialized'].sel(lead=l).swap_dims({'init':'valid_time'}).convert_calendar('standard',dim='valid_time').swap_dims({'valid_time':'init'}).valid_time for l in hindcast._datasets['initialized'].lead],'lead') hindcast._datasets['observations'] = hindcast._datasets['observations'].convert_calendar('standard',dim='time’)...

NWP examples

1

Example for numerical weather prediction to be added to initialised datasets Data sources (to) implement(ed): - [x] GEFS https://www.ncei.noaa.gov/thredds/catalog/model-gefs-003/202008/20200831/catalog.html - [ ] DWD https://opendata.dwd.de/weather/nwp/ relates to #600

**Is your feature request related to a problem? Please describe.** Less maintenance: - reference compute functions for pm use same inits shortcut

how to mudularize `PredictionEnsemble.bootstrap()`

5

we need bootstrapping to get a measure of significance. but skill we just want to calculate from the original (not resampled) data. we agree that from a computation point of...

currently only correcting `initialized`

[DONT MERGE] add time(init, lead) when initializing PredictionEnsemble

4

# Description add `time(init, lead)` Closes #575 ## To-Do List - [x] create upon instantiation - [x] simplify plot - [x] allow to add as `time` not `validtime` - [x]...