The integration with xarray[0] and the Pangeo[1] community at large are pretty good. Not good enough for sub second response on large datasets but very good for

The integration with xarray[0] and the Pangeo[1] community at large are pretty good. Not good enough for sub second response on large datasets but very good for analytic workloads, especially when you are dealing with larger than memory datasets. If you want fast(ish) performance but you can wait a few minutes or longer its a great tool. Climate scientists love it because it lets them focus on their problem domain instead of focusing on dealing with developing parallel software. If you are using it to magically speedup your api backend you will probably be disappointed, but you will also be just as disappointed trying to use a jackhammer to hammer in a nail. [0] https://xarray.pydata.org/en/stable/ https://xarray.pydata.org/en/stable/ [1] https://pangeo.io/ https://pangeo.io/