Repartitioning and groupby operations on dask efficiently

Question

I\'m quite new to dask and have been trying to read the documentation, etc. but not sure I get how to get it to work efficiently.

I have 16 csv files, ea