The other day I saw an interesting blog post about scatter plots with density quartiles using r.

I liked the idea, but wondered whether a simple kernel density plot would have the same effect. And if not, maybe how difficult it will be to adapt the approach to python. So I created this little jupyter notebook: