I found a potential error in the tests of the certification project “Medical Data Visualizer” (Python Data Analysis Course):
The Test test_line_plot_labels expects a ylabel “total”, but in the example the ylabel is “count”. “count” is also the automatically created when using sns.catplot(kind=“count”) and IMO the scientifically more precise one.
As a side note, I found another issue with this:
In the draw_cat_plot function the comments say to split the data by “cardio”, and the test_bar_plot_number_of_bars Test also assumes, that you split the data and made two subplots (or more specifically the CatPlotTestCase.setUp function does in line 10). However this ignores the far easier way of not splitting but specifying sns.catplot(…, col=“cardio”), which achieves the same image in a single subplot. But then the tests fail, as fig.axes is now an np.ndarray of multiple axes. I understand this is hard to test against, but I would make these assumptions where multiple ways are possible explicit in the exercise.
Anyway, these are great exercises, so also thank you for creating them. (And going through the test-modules in these to see whats wrong is also a great learning experience, although sometimes a bit frustating ;))