Add interactive example with Binder

[matthewfeickert]

This Issue is migrated from NEONScience/NEON-utilities Issue #43.

[matthewfeickert]

@cklunch @chrlaney To get the "Use the neonUtilities Package to Access NEON Data" example usefully interactive in Binder the specific data used in the example is needed. In the Rmd though it mentions that

To stack data from the Portal, first download the data of interest from the NEON Data Portal

however data isn't introduced until later in the "Download files to be stacked: zipsByProduct()" section. Is the NEON_temp-air-single.zip data that you used in the first part of the example publicly available? If so, it would be nice to have to have the tutorial be fully interactive from Binder.

[cklunch]

Many of our users download data from the Data Portal, instead of using the API or zipsByProduct(), and the tutorial is intended to serve those users as well. If you download any amount of Single aspirated air temperature data from the portal, you'll get a zip file with the appropriate name. But if you prefer to leave out the portal-downloaded section in the interactive version, I think that would be fine too.

[matthewfeickert]

If you download any amount of Single aspirated air temperature data from the portal, you'll get a zip file with the appropriate name.

Okay great. The example doesn't say where this NEON_temp-air-single.zip came from (it just is referenced) so downloading with wget can just be added into the example.

But if you prefer to leave out the portal-downloaded section in the interactive version, I think that would be fine too.

No, showing multiple ways to get data is useful and since any example should be runnable from start to finish that will be important to keep in.

At the moment, in PR #406 I'm just using jupytext to turn the Rmd that you wrote into a Jupyter notebook. However, given the way that knit treats the formatting of stdout it defaults to an executable Jupyter cell (c.f. the binder link in the comment) rather as rendered output. In your mind does this make the notebook in Binder look confusing? If so, there can just be a corresponding Jupyter notebook created which effectively ports the example from the Rmd, but that looses the benefit of having the example exist as a single source file.

[cklunch]

Thanks so much for putting this together!

I see what you mean, it's a little confusing to have the output render as an executable cell. Is there a change I could make in the Rmd that would fix the problem? We have some standards we have to follow to get the correct rendering on NEON Data Skills, but if it's possible to make a single source file that works for both purposes, I'm very happy to make the changes.