JOSS review: notebooks

[znicholls]

• https://github.com/JGCRI/stitches/blob/main/notebooks/stitches-quickstart.ipynb runs nicely, there are a few tweaks which could help I think:

  • explain in a bit more detail what exactly the get_chunk_info and chunk_ts functions are doing
  • show the dataframes that are being used (e.g. what does my recipes look like, archive data etc.)
  • explain how the archive data was produced in the first place (this seems fairly key and I think an example would really help)
  • variable names could be slightly clearer e.g. data doesn't really tell the user anything about what the variable is
  • the diagram is super nice, well done on that

• is https://github.com/JGCRI/stitches/blob/main/notebooks/full_pipeline_example.ipynb meant to be for users or is that just internal?

openjournals/joss-reviews#5525

[znicholls]

Comments from second round of review:

• a script or docs or something that shows how CMIP data is pre-processed, particularly details that are easy to get wrong e.g. area weighting, would be very useful still I think. Yes, this pre-processing is not necessarily part of stitches but it is important to get right in order to use stitches correctly so I think more discussion of this pre-processing step would be worthwhile to give users a better chance of using stitches correctly.

• install_package_data seems to download all the data every time, even if data is already downloaded. Is that intended?

• I would put a link to the ESD paper in the notebook where it is mentioned (at the moment it is just referred to but no link is included)

[abigailsnyder]

a script or docs or something that shows how CMIP data is pre-processed, particularly details that are easy to get wrong e.g. area weighting, would be very useful still I think. Yes, this pre-processing is not necessarily part of stitches but it is important to get right in order to use stitches correctly so I think more discussion of this pre-processing step would be worthwhile to give users a better chance of using stitches correctly.

• this has been added https://github.com/JGCRI/stitches/blob/main/notebooks/preparing-input-data.ipynb

install_package_data seems to download all the data every time, even if data is already downloaded. Is that intended?

• a note that this is the case has been added to the quickstart notebook

I would put a link to the ESD paper in the notebook where it is mentioned (at the moment it is just referred to but no link is included)

• this has been done