The-turing-way: Blog post: Pros, cons and best practices of using Jupyter Notebooks

[KirstieJane]

This issue is to capture the good points and the bad points of jupyter notebooks.

• debugging and version control is hard
• could you be doing this elsewhere?
• need to run in a set order (not always done!)

Talk at jupyter con: https://conferences.oreilly.com/jupyter/jup-ny/public/schedule/detail/68282

Nice contrast to perl (spelling?) in Rmarkdown that converts Rmarkdown files to scripts for more generalised development.

Important to note that binder doesn't require you to use jupyter notebooks - task for @sgibson91 to highlight how to use binder with other (better?) ways of working.

[sgibson91]

5 reasons why jupyter notebooks suck

[annakrystalli]

The First Notebook War _So Joel Grus doesn't like Jupyter notebooks. Here are some of my thoughts on notebooks, IDE, and R Markdown_: Response by Yihui Xie (knitr developer).

[alexmorley]

I have many opinions on this which I will mostly keep to myself but... I would like to propose a different way of framing this.

Not: why are notebooks good or bad.

But: when are notebooks the best tool for the job and when are they not.

[KirstieJane]

:100: @alexmorley - that did come up in our conversation in person but I don't think I captured it all that well in the issue. I completely agree that an angle of "choose the right tool/language etc" is a great angle to focus on 😸

[sgibson91]

@iamleeg and I have been discussing pros/cons/best practices of using jupyter notebooks in data science/research and would like to write a blog post. Then I had the brilliant idea of putting this blog post in The Turing Way! So if it's ok, we might hijack this issue to plan that post 😄

[sgibson91]

@iamleeg I think we should go down the above proposed route of "when is it appropriate to use a Jupyter Notebook for research", note what notebooks do well and what features might catch people out, and then compile a list of useful extensions that solve (or attempt to solve) those gotchas. What do you think?

[iamleeg]

@sgibson91 Yes, that thing! "Here are some justified good practices when using Jupyter for research" is better than "yah notebooks suck urrr"