Create notebook for EDS book submission

[abnerbog]

This PR is to add content to the notebook for the EDS book submission as described in the R notebook repository template provided by the EDS book publishers.

[github-actions]

👈 Test this PR on Binder

[abnerbog]

The pipeline now runs in Binder!

[lindsayplatt]

I tested this on Binder and it was pretty smooth👍 Some overall comments are below

• In the title and opening description, lean into this being a water focused example. This is our hook for people, so i think leaving out targets could even be OK. Maybe "Using a robust data pipelining tool in R to build a reproducible hurricane data visualization with multi-agency water data"
• I think the purpose should really lean into targets being a tool that can be used in a lot of applications to improve robustness and reproducibility. We are just showcasing its power with a hurricane example.
• Nitpicky, but USGS spells it "gage". So when referring to their system, that's what I would suggest using.
• Add a note after the ESIP one saying "and recreating a USGS visualization by leveraging their open-source code."
• do I need to check in _targets folder to GH for this content to render properly in Jupyter notebook? Hmmm I don't know. What do you mean by "render properly"? When I launched Binder to review the notebook, I was able to run that tar_visnetwork() function without an issue. I did have to install visNetwork first and then run the target twice to see the visual output. That function opens in "Viewer", so that might be why you can't see it in the notebook.
• My overall feedback is to just really make sure the narrative text is emphasizing the linkages and rebuild capabilities that come with building analyses in a targets pipeline. It might be worth talking through how this could all be written in one script, but then changes may require you to rebuild and redownload everything?
• If allowed, linking to our ESIP blog post at the end could be nice.

[abnerbog]

Thanks for this review Lindsay! Will incorporate these suggestions, merge the PR and will reach out to Elise later today with a link to the Binder notebook -- I have a personal computer at my house that I will use to confirm that the notebook runs properly on another machine and OS before merging the PR.

Around this comment:

do I need to check in _targets folder to GH for this content to render properly in Jupyter notebook? Hmmm I don't know. What do you mean by "render properly"? When I launched Binder to review the notebook, I was able to run that tar_visnetwork() function without an issue. I did have to install visNetwork first and then run the target twice to see the visual output. That function opens in "Viewer", so that might be why you can't see it in the notebook.

What I was getting at here is probably more of a question that I should confirm with Tony/Irene around best practices for sharing Jupyter notebooks with the community. I have noticed that some of our notebooks we help create include all the cell outputs (example) and others don't (example).

The notebook that I pushed includes the tar_make output but does not include the tar_read outputs which I thought would be helpful for quickly understanding the material without needing to run any of the code. I was wondering if the _targets folder would need to be checked into GH to display the tar_read outputs when the notebook is checked into GH and first opened in Binder.

For now I won't check in the _targets folder due to time but was just something I was thinking through to ensure the content is easy to understand, review and share.

[lindsayplatt]

Ah yes, I see the issue there. You're right -- there may be a limitation with needing the _targets folder to already be built to see any tar_read() outputs.

[abnerbog]

Just dumping a thought here for later: it could be helpful to modify @lindsayplatt's visual (probably more high-level with some icons) to illustrate the dilemma with creating complex environmental data workflows:

[abnerbog]

Merging this PR now!