I think because you'd want a stream of user info for your machine learning application.

No API, have to be logged in to view, then implement anti-scraping measures.

Because the data tree isn't available for free.