Mardi 19 novembre 2019, à 14h, en salle 26-00/332, Jussieu
During recent years, the study of population dynamics from mobile traffic data has proven to offer rich insights into human mobility laws, disaster recovery, infective disease epidemics, commuting patterns, urban planning, measurement of air pollution in cities, and measurement of energy consumption of cities. These studies have demonstrated how data collected by mobile network operators can effectively complement, or even replace, traditional sources of demographic data, such as censuses and surveys. We present here a series of works that we developed method to extract mobility information from mobile phone data aim for public transport authorities. With the first study, we developed an unsupervised algorithm that enables the mapping of mobile phone traces over a multimodal transport network. One of the main strengths of our work was its capability to map noisy sparse cellular multimodal trajectories over a multilayer transportation network where the layers have different physical properties and not only to map trajectories associated with a single layer. In a second study, we proposed a new approach to infer population density at urban scales, based on aggregated mobile network traffic metadata. Our approach allowed estimating both static and dynamic populations, achieved a significant improvement in terms of accuracy with respect to state-of-the-art solutions in the literature and was validated on different city scenarios.