It's not in these notes, but you can do some super cool stuff with the linear algebra of polynomials. Using inner products, you can not only rederive linear regression, but also its polynomial generalizations, and even zany things like finding the best fit of the form ax^3+cx (missing x^2 and constant terms).