Knowledge Discovery from car sharing data for traffic flows estimation

The newly introduced car sharing services are an unexploited source of data that could be used to estimate the state of the road network as well as to provide new interesting analysis on urban mobility. In this paper we propose a Knowledge Discovery System that first gathers information from car sharing sites and applications, and then processes it to estimate interesting metrics such as travel time and vehicle flows in the urban areas at different times and in different days. We further argue that the information gathered can be processed in real-time, to estimate instant traffic, and can be exploited to perform deeper analysis, using historical data. Finally, we analyze vehicle availability as a function of time in different zones and show how the results can be applied to travel time estimation, car stockout forecast and multimodal travel planning.