parallelize data analysis with dask

[rabernat]

once #1, #2, #3, and #4 are done, we can think about parallelizing the analysis with a dask cluster. Basically, the idea is

• create a list or all the files to process
• define the algorithm to run + some reduction operation
• feed it to dask and let it do its thing