SciPy Tutorials 2018

This repository organizes efforts to use Dask within tutorials for the SciPy 2018 conference. Our objective is to complement existing tutorials with a small Dask section that can show how to parallelize the algorithms and APIs presented within the tutorial.

We expect that doing this effectively will require collaboration between a few groups:

1. The organizers of the tutorials who understand the domain, algorithms involved, opportunities for parallelism, and how to communicate effectively to the audience
2. Developers with Dask expertise who may be able to help tutorial organizers develop parallel computing examples in some cases
3. Developers with JupyterHub or cloud deployment expertise who can help to ensure that different environments are publicly available and can help triage issues with load during the conference.

Example Infrastructure

Initially we are considering modeling this after the pangeo.pydata.org deployment that combines JupyterHub, Dask, and cloud data storage. This deployment includes software and examples for the earth science community (NetCDF, XArray, ...) and has had broad impact in advancing the discussion of the future of scalable computing within that domain. We encourage those interested to try running examples on that infrastructure to get a sense for what a tutorial experience might look like. Login is publicly available with GitHub credentials.

What to do now?

If you are organizing one of the SciPy 2018 tutorials we encourage you to raise an issue with the title of your tutorial. The issue template has a few questions to start discussion.

Managing Expectations

This is intended as a collaboration between the three parties above. However we expect that most of the content generation will be handled by the tutorial authors themselves.