Synthetic Household Travel Data Using Consumer and Mobile Phone Data

This research develops a method that fuses consumer marketing data with anonymous, passive location data to create synthetic populations with individual-level synthetic travel diaries. The travel diaries detail each person’s travel and activities at locations in a timeline format. This low cost synthetic data could give entities up-to-date, detailed data that match the population’s short-term and long-term movements. The research team previously built an initial implementation of the data fusion process in Atlanta, Georgia. This IDEA project proposed building a synthetic household travel dataset for a different city that had a larger study area in land and population size. The four-county planning region of the Puget Sound Regional Council in metropolitan Seattle, Washington was selected. The project aimed to test the transferability of the concept, the scalability of the method to larger study areas, and the suitability of the specific implementation method chosen in the initial study. The research focused on developing a process that will be consistent nationally, rapidly deployable for any size city, and systematically updateable over regular time periods. The first stage of the research effort refactored the methodology from the initial study in Atlanta so that the same code would build a synthetic population and travel diaries in the larger metro region of Atlanta and in metro Seattle. The required “big” data were obtained for metro Seattle and the synthetic travel diaries were built. The second phase of the research effort validated the resulting synthetic travel diaries for both Atlanta and Seattle against external sources. It also checked for internal consistencies with the passive data that were fed into the synthesizing method. Based on implementations in Seattle, Atlanta, and Asheville, North Carolina, and on the comments by the project's expert panel, it was concluded that this method would be most useful, as it is functioning right now, in small-and medium-sized regions for planning. For large regions that have invested in large household travel survey collection programs and sophisticated activity-based models (ABMs), more research is needed to merge a data- driven approach into existing activity- based models. For state departments of transportation, this data-driven approach lessens (or removes) the need for a statewide household travel survey program, standardizes travel models in use without time-intensive local calibration, and standardizes analysis of projects for transportation improvement programs within the state. For regions interested in analyzing the impact of autonomous vehicles (AVs), this method can be combined with open-source MATSim to rapidly analyze short–term responses to AVs assuming shared fleets, privately owned fleets, or a mix of the two.


  • English

Media Info

  • Media Type: Digital/other
  • Edition: Final Report
  • Features: Appendices; Figures; Tables;
  • Pagination: 31p
  • Serial:
    • Issue Number: 184

Subject/Index Terms

Filing Info

  • Accession Number: 01643631
  • Record Type: Publication
  • Report/Paper Numbers: NCHRP IDEA Project 184
  • Files: TRIS, TRB, ATRI
  • Created Date: Aug 11 2017 2:00PM