Python port of the salary data example

[chisingh]

This is a basic working example with the salary data to generate plots inside Jupyter notebooks. Closes #104

[jhofman]

@sharlagelfand will have a look over, @jhofman will try to clone and render the notebook for reproducibility

[sharlagelfand]

Thanks @chisingh! Just some initial thoughts here, mainly around consistency and how the python code fits within the existing repo (which follows the structure of an R package, for now):

• Could we rename the datamations folder to e.g. python, and everything would be self-contained in there? Then, if we decide to put all the R code into a folder, it could just go in e.g. R and follow the same convention, plus it's clearer from looking at the main repo what it contains.
• Because the current structure is that of an R package, I can ignore entire folders in the R package building process (e.g. sandbox is ignored) but it gets a bit complicated if things are mixed, e.g. the python tests living within the same folder as the R tests. Could the python tests live within the python folder?
• Related, can the test python specs live in the python folder, e.g. in python/tests/test_specs or something like that, so it's clear that these are specs used for formal testing rather than sandbox, which is more for playing around while working out new features.
• It would be nice if the python README contained a rendered example, e.g. the same way the main README does - these are just screen captures of the resulting animation, but it's nice to see that things "work the same" without running the notebook myself!
• The main function name is not consistent between R and Python - datamation_sanddance() versus datamation() in Python - I assume we want to stick with the same one

I think that's it for now! Excited by your demos of this and to start testing parity 🎉

[chisingh]

Moved python tests to the same folder as the rest of the source code and fixed the README and the function name datamation_sanddance. The folder name datamations also becomes the python package name, so let's keep it for now as you can ignore it entirely from the R package.

[sharlagelfand]

Thanks - looks good to me, then! Understood re: the folder name becoming the package, all good there then :) I'll push a commit to the main branch that'll ignore the datamations folder so it won't cause issues on the R package checks.

[chisingh]

Thank you!