There are some books here, but all focus on linear models.
Do we have books/notes with broader coverages?
In case of linear models, “coordinate-free” basically means “matrix-free” and uses the theory of vector space, and the viewpoint is geometric. (Outside of linearity, would we go into something like Information Geometry?)
Edit: It seems Information Geometry and Topological Data Analysis are related to this question. But books on IG and TDA tend to collect recent research results, and the problems solved are specialized. Rather, I’d like to see coordinate-free reconstructions of main stream statistical methods (maybe also advanced ones), and this would also help to learn IG and TDA, proper.