[ ] add old data like chocolates, spam for exercises
[ ] at least, need some simpler classification data sets,
[ ] perhaps some simulated to explain what non-linear or separated mean in high-d, expand the intro to classification to include multiple boundaries possibility, and maybe large p?