Add a very simple analysis in your CI

[HealthyPear]

You could add (an updated version of) the PG 1553 tutorial and compare the results against some stable run in order to catch easily issues like #566

if that test does not pass the deploy workflow should not be allowed to run, preventing broken releases

[HealthyPear]

You could either test the produced images against reference ones (the dataset is fixed and you analyze it with the exact same config)

https://github.com/matplotlib/pytest-mpl