The top thing is the difference in funding, in both time and what you have to do to get the money. 12-month funding is very different from 9-month funding: with 9-month funding you have to scrabble around during spring semester trying to figure out how you are going to pay your rent and feed yourself over the summer. With 12-month funding, you do not. Secondly, an RA is way better than a TA - TAships are quite time-consuming and you will still be doing 20+ hours of research a week anyway, especially if you want to go into academia. RAships pay you to do things you would already be doing, and you can always TA some classes if you want to later (potentially for extra pay).
So for that reason alone I would select UNC.
Both are good schools, highly ranked in statistics. UNC probably has the edge in academic circles overall, and of course their SPH is ranked #2. Biostatistics doesn’t limit you to work in biomedical and health sciences; you learn all the same models and stuff, you just apply it to
Other thoughts:
-Chapel Hill is a quintessential college town; Raleigh is a medium-sized Southern city. They are within 40 minutes of each other. Raleigh is pretty safe; crime is far below the national and the North Carolina average. Chapel Hill is pretty well culturally-developed despite being a small city because if its proximity to Raleigh and the presence of the university. I really think you can’t go wrong here, and living in one you can make frequent trips to the other. Pretty sure you’d need a car in either.
-Statistics is having a moment right now, a moment I am not entirely sure will be over in the next 5-6 years. As more and more companies implement electronic systems that collect large amounts of data, more of them are clambering to find statisticians and data scientists that can help them organize, analyze, and interpret that data and use it to do something. So the industry opportunities are very, very good in this field - particularly if you have at least basic to intermediate computer programming skills. If you can do a certificate program in computer science as a doctoral student at either school, I would do that (and honestly, the presence of such a program would give that university an edge for me). If not, at least take some classes when you can. Take a look at some recent job ads for statisticians and data scientists and look at what they ask for, then develop those skills through coursework and other options like consulting. (DEFINITELY do consulting. I think most academic programs in stats require consulting hours anyway, but do more than required. I did some paid stats consulting in grad school and it was absolutely instrumental in getting hired. Plus you make $$$, which is always a good thing.)
You could also straddle the two and go nonprofit/think tank. Either of the two universities you want are EXCELLENT for that because of the proximity to the Research Triangle, especially RTI. There are lots of professors at both universities that work at either.
AS far as academia goes I think opportunities in statistics are pretty robust because of the competition from industry.