For the CIROH User's Conference in May, a collection of domain scientists will be congregating to learn about CIROH/OWP tools for their research. Having some publicly available MVP tools will be useful.

Some of the capabilities we should target are:

1. Public repo.
2. PyPi package public and pip install ready.
3. Build image(s) on some public repository.
4. Live documentation.
5. At least one user tutorials with examples. #53
6. High test coverage (XX%+)
7. GitHub Actions for pytest, pytest coverage, linting, and docker build and banners for test status and coverage.
8. Obtained user feedback. Related to #53.
9. Some functioning MVP:
   • Raster categorical
   • Define public API
     • High-level public API: agreement_map, contingency_table, metric_table = xr.gval.categorical_compare()
     • Second tier public API: See issue #53
   • Compute agreement maps
   • Compute contingency table
   • Crosstab data
   • Compute metrics