I don't know how much you can tell about a city's drivers by walking around. Probably not much. I remember walking in San Francisco in 2000 and being very conscious of the fact that I had seen reports on the local news saying that the city had a poor pedestrian safety record. But of course I was a pedestrian.

Last year I spent a few months in Washington, D.C. and as a pedestrian I felt fine. But the driving behavior I saw at a few places - including two intersections in particular near the Takoma Metro station that I walked by every day on my way to and from the place I was staying - looked pretty dangerous. People would race down two lane streets to try to be the first to reach the spot where the street collapsed into a single lane; people would pass on one-lane streets by pretending to wait to turn right and then slamming down on the gas to go forward as soon as the light changed; people would cut off oncoming traffic to turn left when lights changed; and in general people did not leave much time to make turns. Had these been a few isolated incidents here and there, I wouldn't have thought much about it, but I regularly found myself thinking "that was insane!" after witnessing something crazy from the sidewalk. But I never saw an accident.

Recently I saw a story in the Examiner (via somewhere I don't remember): "D.C. drivers most accident-prone in nation, insurance study finds":

D.C. drivers are more likely to be in auto accidents than drivers in any other city in the country, and Alexandria and Arlington drivers follow closely behind, according to a new study.

D.C. drivers average one accident every 5.4 years, making them almost three times more collision-prone that drivers in Sioux Falls, S.D., which ranked as the safest driving city in the 2008 Allstate America’s Best Drivers report.

The number means D.C. drivers are 84 percent more likely to be in an accident than the average driver nationally and places the city as the most dangerous for drivers among the 193 studied.

The rankings can be found here (under 2008 data). Note that San Francisco also does poorly in the rankings at 185, but there's a huge gap between D.C. and almost all of the other listed cities in terms of "Relative Collision Likelihood" (compared to the national average): D.C. comes in at nearly 84% more likely and S.F. at just over 44% more likely. I haven't looked into the methodology, but the Examiner says that the numbers are based on claims filed in 2006 - so that's before I went to D.C.

That Examiner article also provides some information about congestion:

The Washington area ranks as the second-most-congested in the country, tying with Atlanta and San Francisco and trailing only Los Angeles.

The interesting thing there is that the Washington area, Atlanta, and San Francisco (assuming this means the San Francisco area) are the three regions that built similar heavy-rail transit lines after the second world war. I suppose a transit skeptic would take this as evidence that the rail lines haven't done their job, but the fact that Los Angeles, which did not build any lines back then ranks lower, suggests that the costs of not building transit could have been higher. It also suggests that the transit projects facilitated growth (and at least in the case of BART, that was one of the goals).