Urban Traffic Modelling and Prediction Using Large Scale Taxi GPS Traces

Monitoring, predicting and understanding tra c conditions in a city is an important problem for city planning and environmental monitoring. GPS-equipped taxis can be viewed as pervasive sensors and the large-scale digital traces produced allow us to have a unique view of the underlying dynamics of a city’s road network. In this paper, we propose a method to construct a model of tra c density based on large scale taxi traces. This model can be used to predict future tra c conditions and estimate the e↵ect of emissions on the city’s air quality. We argue that considering tra c density on its own is insu cient for a deep understanding of the underlying tra c dynamics, and hence propose a novel method for automatically determining the capacity of each road segment. We evaluate our methods on a large scale database of taxi GPS logs and demonstrate their outstanding performance.