"safety/danger per-mile" is not a very reasonable metric in many cases though.
E.g., for airplanes, most of the danger is in takeoff/landing, and whether you fly 300 miles or 10000 miles in between is much less significant. For cars, long distances often involve highway driving, which is safer, and driving to the airport probably(!) doesn't involve alcohol, which is a much bigger factor in overall accident rates... etc etc.
What would be more interesting to know is the risk per flight, and compare that with a risk estimate for a given airport drive...
[In my case I just take the train to the airport from my house, and the lines I use have fatality rates which are essentially negligible (1-2 per century). I suppose I'm mostly likely to get hit by a car on the walk to the station... or just slip on the steps. >< ]