Where bike lanes went in, car traffic fell — but crashes per car did not.
Bike lane mileage tracked against car traffic and crash rates across 18 cities. Traffic fell where lanes went in; crashes per car did not.
Not real data. The numbers on this page are invented. They were generated by a script for a training exercise and describe no real city.
Average fall in car traffic
5.3%
across 18 cities, from 1.5% to 10.2%
Average change in crashes per car
+0.1%
crashes per 10,000 daily car trips, before against after
Cities where the crash rate fell
9 of 18
close to the half you would expect from chance alone
Car traffic decline after new bike lanes, by city (%)
Pittsburgh10.2
Boston9.4
Albuquerque9.2
Minneapolis9.0
Columbus7.6
Oklahoma City6.3
Austin6.3
Nashville5.7
Portland5.7
Milwaukee4.5
Denver4.5
Tucson3.2
Fresno2.9
Seattle2.9
Sacramento2.5
Raleigh2.1
Omaha1.6
Louisville1.5
More lane, less traffic
Each dot is one city. Cities that built more lane saw larger falls in car
traffic (correlation 0.94). A correlation this tight is a property of
invented data — real cities are messier, and a relationship like this one
still would not prove the lanes caused the fall.
But the crash rate stayed put
BeforeAfter
Crashes per 10,000 daily car trips, before and after, with the change at
the right. The pairs barely separate: the rate fell in
9 of 18 cities, which is what chance alone would produce.
Fewer cars meant fewer crashes in total, but no city got safer per car.