When I visited one of the universities I got in to, I was given a piece of friendly advice:
If you really, really love somewhere, head over heels, completely smitten - consider not doing a PhD there. It is very rare that someone will get a faculy position i nthe same place where they did the PhD, so if you have two options for grad school that are both very good, but one you are in love with in terms of location and other things that may last a while, go to the other one.
Also, here are some things I am considering (in addition to above, maybe, I'm not convinced), which were not mentioned:
- who wants to go here? is this a name-brand school, where people may want to go because it's OMG, ___! ? i do not want this: many students there are there for the wrong reasons, and could have had a better match at a lower-ranked uni. of course there are also the brilliant students, but a fair number of brilliant students are just people who lucked out with undergrad research advisors and have grown an enormous ego. also not compelling to me.
- who else lives here? for example, consider California. I'm not a resident, and I don't like the weather. But a lot of people want to be in california, and the cost of living is quite high. i am choosing to say no to all california schools simply because i don't like the prospect of high living expenses for being surrounded by people motivated by totally different reasons than me.
- what about babies? i am planning to get married, and i don't want to rule out having/adopting children over the next half-decade. obviously, a reasonably family-friendly environment is good. since i am a female in cs, this means that i am looking for evidence in each school of successfully graduates students who had or adopted children when both they and their spouse were in school or working, and how long ago that was.
These are a bit different. Obviously I am considering all the other factors listed in this thread.