Hacker News Reader

Does this also work for larger numbers of features? (like, 2000, as opposed to 6-9 in the demo)

Only if your dataset is really small. It only supports up to a low millions of points