CMU vs Columbia Statistics PhD

[grabianco]

What do people here think about CMU vs Columbia Stats PhD? I'm mostly interested in machine learning and the statistical theory behind it and both programs have faculty I'd be really interested in working with. I'm not 100% sure if I want to go into industry or academia, but I'm slightly leaning towards academia right now.

Both locations' geographies have their advantages and disadvantages. Pittsburgh seems much cheaper which is an important consideration for me, but also has bad weather I heard. Meanwhile, I like the big NY city environment having grown up in one myself, but I know it's ridiculously expensive.

In terms of the programs themselves, It seems CMU is a slightly more prestigious program (in terms of rankings and in terms of their CS dept), but Columbia might be a better name for positions in industry? 

If anyone has any useful input about either the programs, or the schools or locations in general, that would be greatly appreciated! I'm trying to make this decision as soon as possible since there are presumably people on the waitlist at both programs.