To help cities figure out where they need roads, transit, buildings, sidewalks, and other infrastructure, transportation and urban planners must understand three components of mobility:
- households and the people in them,
- firms and their employees and customers, and
- the movements between them.
Traditionally, planners collect information about these using small-sample household surveys. Then travel demand models expand the sample to represent a whole city and to forecast each of the components above given different futures. Unfortunately, the surveys used to build today's travel demand models are expensive, infrequent, and onerous to the respondents. Each of these problems can reduce the data's usefulness and bias the resulting model.
At Transport Foundry, we have been working hard to develop a new approach to quantify mobility using passive data. Instead of actively collecting survey information from a small sample, we use passive data collected for a large proportion of the population during their normal activities and travel without any overt interaction. These data capture individual preferences and behavior, such as location data from mobile device carriers and household information from commercial marketing.
Our approach fuses passive data—including data on households, firms, and regional travel—with the National Household Travel Survey in a way that the data and the model become one and the same: a data-driven model. Because of the large sample sizes available in passive data (generally over 30%), our travel demand model avoids the statistical gymnastics necessary to expand a <1% survey sample to represent an entire city. Rather, our model is consistent nationally, rapidly deployable for any size city, and frequently updated with fresh data. Most importantly, the rapid deployment schedule gives planners more time to solve their cities' transportation problems.
To date, we have applied the data-driven travel model in three cities: Atlanta, Seattle, and Asheville. This work began with funding from the National Science Foundation and the Transportation Research Board. The TRB IDEA Program technical report summarizes the approach and its validation.
Over the next few weeks, we will walk you through this report, diving deeply into what it means for transportation planning. Watch the CityCast category for more.