In Traffic: Why We Drive the Way We Do (and What It Says About Us), the author states that commute times throughout history have remained steady at roughly a half hour in each direction. Advances in transportation technology (our feet, horses, bicycles, trains, automobiles, flying cars, etc.) allow us to live farther from where we work. This got me thinking about my own commute from Berkeley to San Francisco, how it compares to those of my neighbors, and how commutes vary across the country.
This visualization draws a red circle at the center of each zip code indicating the number of people who travel to or from that area. The blue lines show the driving directions.
The primary data source is the CTPP 2000. This survey was sent to a subset of households during the 2000 Census and records, among other things, where people live and work.
The CTPP has since been superseded by the ACS. Although transit statistics from the ACS have been published more recently, the the new data is not sufficiently fine-grained for use with this map.
The CTPP provides data on a census tract level. However, this map uses zip codes to identify regions because they're more familiar to most people. The mapping from census tract to zip code is done by using the Census Zip Code Tabulation Areas (ZCTAs) to determine the proportion of each census tract that falls within each zip code and weighting the CTPP data accordingly.
The Google Maps API is used to determine routes and transit times. The usual caveats apply. In particular, it is assumed that all commuters drive during non-peak hours. This is surely incorrect (but it's the best that I can do) so the trip times are likely to be underestimates.
Although an effort was made to reflect the data accurately within the constraints of the design, correctness is not guaranteed. If I made a mistake, don't shoot me! Let me know and I'll see if it can be fixed.