How does one go about choosing a graduate school for computer science?

Look at what they research. It will more or less define any research you do. If you don't do research it will color what your professors know, and how well it is covered in your classes. Another is how well is the school ranked, generally better ranked schools attract better candidates and better professors, cooler research etc. For instance consider DARPA grand challenge entries. How many were done by top universities and how many were done by community colleges or even universities ranked middle of the pack?

And last but not least soft issues, cost, location, etc.

I'm assuming that you have an idea of what interests you and that you are shooting for a PhD or masters by research.

Research your topic of interest on Google Scholar. Read widely to identify who are doing comparable work.

Write a research proposal and send to the professors who are doing and supervising such work. You need to be very passionate about your topic in order to sustain you through 3-5 years of grad school. Of course, what you write your thesis on won't resemble your research proposal. That's just the way it is. You discover so much as you dig deeper and deeper into your chosen area.