Hey @tino-michael

This is definitely possible, it's a matter of finding someone with time to implement, test, and document it. I'd be happy to review a PR if you can take this on.

In the meantime, I'm tempted to simply kick this over to seaborn, which already handles this case nicely

As far as I can see, seaborn also "only" allows to print the quartiles with the inner parameter but not user-defined percentiles (some colleagues insist to see the "1sigma" -- i.e. 68th percentile -- point in the distributions).

I will try to find some time (you are right on that regard) to dig into it. I just thought that it would be quickly enough to implement for somebody that is already familiar with the code -- which I am not.

Hello,
What color should the percentile lines be? And what size in comparison to the mean/median lines?