Seminar: Improving the Quality of Urban Public Transport: Lessons Learned from a Novel Use of GPS Data

In Italy, public transport is typically operated by private companies under service contracts including penalties for non-compliance. Therefore, detecting anomalous traffic days is key to fairly assigning delay responsibility for buses operating in mixed traffic: operators should not be penalized on days of exceptional traffic conditions. To address this, harmonic average speeds of vehicles are used to cluster hours of the day exhibiting similar traffic patterns. For any of such traffic states, a Generalized Additive Model is fit to the available speed percentiles in order to detect anomalous days. A case study relying on GPS data is presented for the city of Trieste.

Presented by Andrea Mecchina, PhD candidate, University of Trieste, Italy