SciPy 2025 tutorial plan

[TomNicholas]

As discussed yesterday - here's a meta-issue for discussing tutorial content for this year. This is the schedule we currently have, but we can and should make edits / replacements as necessary. Please feel free to make edits, create sub-issues, ask questions etc.

• (10m) Introductions
• (1hr) Understand xarray’s core data structures (Ian presenting, Ian preparing, support from Tom, Justus, Eni)
  • (30m) Lecture
    • (10m) Named arrays and coordinates (Variable)
    • (10m) Groups of arrays with coordinates (DataArray and Dataset)
    • (10m) Hierarchical trees of related groups (DataTree)
  • (20m) Exercises
  • (10m) Break
• (1hr) Understand how to map typical xarray computations and workflows over hierarchical data (Justus presenting, Justus preparing, support from Alfonso + Tom)
  • (25m) Lecture
    • Non-datatree computations
    • Datatree computations (datatree computation #328)
  • (25m) Exercises
  • (10m) Break
• (1hr) Hierarchical storage formats and their data models (Tom presenting, Tom preparing) - Page on storage formats #325
  • Could just rip this whole section from the cloud-optimized geospatial formats guide?
  • (30m) Lecture
    • (10m) HDF5 (including differences to DataTree)
    • (10m) Zarr (including differences to DataTree)
    • (10m) TIFF
  • (20m) Exercises
  • (10m) Break
• (1hr) Intro to Zarr on the Cloud (Joe presenting, Negin preparing, support from Joe + Tom) - Zarr tutorial #323
  • (30m) Lecture
    • (10m) What is Zarr? (@rabernat said he already has a notebook for this somewhere?)
    • (10m) Object storage as a Zarr Store
    • (10m) Accessing open data as Zarr on the cloud
  • (20m) Exercises
• Open a provided Zarr store on the cloud
  • We decided to use an existing non-DataTree example for expediency: ERA5, but...
  • I just noticed we already have a section for this, on CMIP6!!

We also need a few small (MBs or less) example datatree tutorial datasets to use for some of this content. Ideas are here, but summarized as:

• Radar (Alfonso) - Adding hierarchical radar data pydata/xarray-data#35
• Satellite imagery (Joe)
  • Sentinel EOPF?
• CMIP6 model ensemble (Alfonso + Tom)
• Bio (Ian)
  • OME-TIFF (https://openslide.cs.cmu.edu/download/openslide-testdata/Generic-TIFF/)
  • Plate inheritance?
• GPM IMERG (Eni)

All of these are intended to be smaller examples (<50MB) - they are done once they are in https://github.com/pydata/xarray-data.

Other miscellaneous tasks:

• Landing page with schedule - Scipy2025 outline page #324
• Contact Nebari people (Ian)

cc @eni-awowale @jhamman @ianhi @aladinor @chuckwondo @negin513 @keewis

[rabernat]

• Intro to Zarr on the Cloud (Joe presenting, Negin preparing, support from Joe + Tom)

On this topic, a Zarr tutorial I gave for OGC in 2022 could be a good starting place:

• Repo: https://github.com/zarr-developers/tutorials
• YouTube recording: https://www.youtube.com/watch?v=unGL07trSjA

[scottyhq]

Hi all! Thanks a million for putting this together, sorry I haven't been able to join planning calls due to being on leave and the timezone, but feel free to ping me on pull requests to review content. Best of luck with it next week!

[negin513]

Hello @TomNicholas ,
Thanks for assigning these tasks. I added some materials for Zarr introduction in this PR #323 and added some information on Zarr-V3 (sharding, compressors, etc). Please feel free to review as I put these together in a few hours. Thanks! 👍

[TomNicholas]

Thank you so much @negin513 ! This looks great.

[negin513]

Hey @TomNicholas @eni-awowale @jhamman @ianhi @aladinor @chuckwondo @keewis @scottyhq ,
Just checking in—how’s everyone feeling about tomorrow’s presentation? Would it be helpful to do a quick sync tomorrow morning to go over anything?

[TomNicholas]

We pulled it off

[dcherian]

Would be good to stick the outline under "Learning Paths" after the dust has settled. Could call it the "Hierarchical Datasets Path"