Columbia's stats program has a great reputation, it's their master's in Statistics that is generally portrayed as not a very good program here. Biostat is completely different department so that doesn't apply at all. UNC has a better program, but Columbia's still a top 10ish biostat department. If you really want to get a PhD, you should look into whether you will have research opportunities at the two schools, but if you do well you should be able to be successful at both. I am an advocate for choosing a place where you will be personally happy - since Columbia has a good department, if you want to live in NYC and money isn't a huge issue, go.