Improvements for ISC 2025!

[danny-lloyd]

• Typos in xarray (e.g. panfas and coorfinates)
• Add xarray.open_dataset to the start of ex01a
• xarray exercise numbering --> change a, b to numbers
• python-intro - we did extra after lists on tuples, sets, dicts as students had programming experience
• xarray ex01b Q5 remove date from question as t is only length 1
• Should we introduce the underlying numpy and matplotlib libraries before cf-python, netcdf & xarray?
  • Alex thinks that rather than change the order, we should ruthlessly remove any reference to the underlying libraries from the high-level presentations (like we do for pandas earlier in, which is also backed by numpy). This gives them easy stuff to use first without causing confusion?
• Fix satellite notebook so it works with Alex's script (tag answer cells).
• xarray plotting exercise - question 7 - answer doesn't match question (2A)
• xarray masking exercise 3b - dataset name is wrong vbhubo.pgc0apr.nc
• xarray - didn't quite get enough time for all xarray exercises (skipped 3a entirely)
• split up cf-python and cf-plot exercises into six separate exercise Notebooks (still corresponding to each of the six sections matching the teaching Notebook, but will mean we don't have to use explicit numbering in the Notebook itself)
• Rename python-data notebooks folder to teaching-materials and add copies of the teaching web pages (as they currently are) to add to the release
• Make a 2024 release, use SingleFile CLI to snapshot all the external docs (e.g. xarray tutorial)
• Formatting: make use of bold text for question numbers and note/success boxes in markdown for info boxes
• We delivered too much material without giving them a chance to do the notebooks at the start of xarray and cf-python, which meant they got a bit behind. I think this is the main issue rather than order etc.
  • We skipped one section of xarray, and just (barely) got through the rest. Would have been good to have a little more time for notebooks here.
  • cf-python was too long and we didn't manage to present all the material.
• Presenting both cf-python and xarray back-to-back DID cause some confusion.
  • A diagram of how all the libraries relate and WHY we are doing more than one would be really helpful.
  • Discussing with participants they did understand why we did this, just took some time to get their heads around.
• python-data - delete the answers from exercise 06b!
• netcdf session material is far too long - but half of it is about do-able and appropriate (in SB's opinion).
  • In most cases they will never have to actually use the netcdf4 library, so really they just need to know it exists!
• We did slicing to death. Drop it from numpy and netcdf.
• Alison's presentation says 'comma separated variables' instead of values
• NetCDF teaching:
  • Q4 final bullet states 'Create two values:' but it should say 'Create two variables:'
  • If the students run a cell and there is an error but some of the previous lines were successful, often a line would create a dimension or variable and then re-running that cell would error on attempting, when re-calling, to create a new such netCDF object with the same name which is forbidden. We should add a note to clarify to the trainees about this and that they should just comment out lines that have already been run to solve it
• Have a look at the satellite notebook - could it be made easier with DataPoint?

[sadielbartholomew]

Good idea @danny-lloyd, thanks for setting this up. FYI I've added one to the list.

[sadielbartholomew]

OK just adding ad-hoc now, as we are all doing! Since the list is getting longer I am attempting to separate into some sub-lists via checkboxes in the second-level nested bullet, e.g. I've put one for the netCDF session/content specifically.