Add timeseries filtering example to illustrate subsetting of columns from selector values

[joelostblom]

I answered this SO question and thought it would be interesting with an example like that there too. The main thing it illustrates is how to select column based on selection values, but also puts it all together in a mini dashboards.

The code is somewhat long, but not long enough to be in a tutorial I think. I chose the "interactive" gallery category, but I think "advanced calc" would also work. There are better ways to compute time series correlation, but that's a parenthesis here.

altair-ex-2024-03-12_10.16.49.mp4

[joelostblom]

The reason for the failing test is that pandas added numeric_only in 1.5 (at least to corr) and in 2.0 the default was changed to False. That means that we need set numeric_only=True as in this example, but in our CI we test the oldest version altair supports which is 0.25. Since the docs build, I believe this should be fine?

[jonmmease]

What are you proposing? I think it's a good idea for the docs to use a more modern version of pandas.

[joelostblom]

I agree and I was initially thinking that we didn't need to do anything as the docbuild ran fine, but I forgot that the examples are part of the test suite too. In my last two commits I fixed the issue I mentioned above as well as a mypy issue with renaming inside reset_index by using a slightly more verbose approach via rename instead. All tests pass now.

[jonmmease]

This looks good! In the future, I wonder if we should split out the ci jobs that run the examples so that we can use more up to date dependency versions. But not something to worry about today.

[joelostblom]

Thank you for reviewing; merging!

[mattijn]

Nice example @joelostblom!

[binste]

Agree, nice one :) And also thanks for the other PRs on improving the docs!