Yeah, that's a really important distinction to keep in mind. I work a lot with crime data, and so people say things like "Oh, so you know all the best places to find drugs", when in reality what I know is the best places to get caught. Where people are committing crimes and where they are getting arrested may or may not be the same.
For San Francisco all the crime data is published (well, other than some sensitive data like the locations of rapes). datasf.org has a listing of all the geospatial data available for SF. A lot of other cities have their own data repositories. Washington DC has one of the best (http://data.dc.gov/) and New York has a good catalog as well (http://www.nyc.gov/html/datamine/html/data/data.shtml).
