Information fusion of smart card data with travel survey

in comparison to traditional travel surveys, smart card transaction data with extensive spatio-temporal richness offers a great opportunity. However the smart card dataset may lack some basic information such as trip destinations and socio-demographic attributes of the traveler. Our goal is to enrich smart card data with this missing information. We propose an information fusion approach of census, travel survey, and smart card data. Our methodology involves solving a maximum weighted bipartite graph consisting of a collection of smart card owners and a collection of smart cards. The links between the nodes are weighted using a trip chain choice model. This methodology is applied on Gatineau's public transit service (Canada). Results are analyzed at macroscopic and mesoscopic level by comparing estimated public transit riders' population with the observed population in the travel survey. Our method produces results that fit the expectation as high as 93%.


  • English

Media Info

  • Pagination: 32p

Subject/Index Terms

Filing Info

  • Accession Number: 01627355
  • Record Type: Publication
  • Source Agency: ARRB
  • Report/Paper Numbers: CIRRELT-2016-49
  • Files: ATRI
  • Created Date: Feb 27 2017 9:46AM