I don't like this chart at all. The data is based on multiyear estimates from the American Community Survey and do not represent a rate at discrete points in time (in this case, 1990, 2000, 2005, 2009). Because the chart presents it like this, and because of the large margin of error, it appears that in 4 years (between 2005-2009) everybody JUMPED whereas it was likely a more subtle increase (although, again because of the margin of error, there could have been a subtle decrease). The margin of error in the ACS dataset is very high because of the low sample. Only a couple thousand people in each of these cities were interviewed yearly and the ACS extrapolates their answers onto the larger municipal population. We need larger sample sizes, but also more detailed insight into transportation. Read more:http://www.stevencanplan.com/tag/american-community-survey/
Interesting. Thanks for the comment, Steve.