Optimizing Dask operations using compute(). isin and shape

Question

I am exploring Dask parallel computation to run lighter on memory and also take advantage of parallel computing making my python code faster.

I am using a 8 core mach