Create an agenda for the tutorial

[negin513]

Since there are different topics/aspects that can be covered for the Dask tutorial and we have a limited time we need to define the scope of the tutorial and create a clear agenda for it.

From our conversation with @dcherian and @vanderwb : This going to be a half-day event with topics and sessions for the following audiences:
(a) folks who are Python-aware but essentially total novices to Dask.
(b) those who use Dask regularly but would like optimization guidance and tips/tricks.

The followings are my initial thoughts on what would be appropriate for this meeting:

• Introductory Dask + Xarray:

  • What is Dask
  • Dask +Xarray
  • dask-backed Xarray objects (lazy computations, actual values)
  • Distributed clusters
  • Extract Dask arrays from Xarray objects and use Dask array directly.
  • Dask Delayed to parallelize any code ( do we have time to include this?)

• Intermediate Topics:

  • Dask chunking schemes (performance and rechunking)
  • Apply unvectorized functions (apply_unfunc)
  • More advanced collection of custom operations (map_blocks, map_partitions, map_overlap do we have enough time for this?)
  • Blockwise computation

Since this is going to be a half-day tutorial, we need to be cautious of the time.

We are going to solicit more feedback on this from the community and ESDS forum.