Clustering

Gist

Who is closer to who?

Types

1. k-means clustering

Criticism

Any use of "data-driven clustering" should be met with some skepticism

1. Frank Harrell seems to be against it.

References

1. https://darrendahly.github.io/post/cluster/ (Interesting blog post about how if you have two correlated variables, clustering can produce a lot of groups. It's WEIRD if it doesn't produce groups)