Comparing multiple data streams to assess free-floating carsharing use

Passive data streams are a great alternative to traditional travel surveys to assess the use and the change in behavior. Because of accessibility issues, researchers have harvested free-floating carsharing (FFcs) web-services in order to estimate the spatiotemporal demand. This paper presents the comparison of multiple data streams in the assessment of trip type distribution for FFcs service. While a full dataset of GPS traces may be considered a good approximation of the ground truth, harvesting of origin-destination data seems to estimate correctly the general trend of certain trip type distributions, while for other trip type estimations, a more extensive set of data is needed (member & stopover information) to fully assess it.