The aim of our work is to extract knowledge from a set of time series. In the context of the ASTRAL project,we want to contribute to the computer science field, but also to the applicative field of urban traffic analysis.
The urban traffic phenomenon comprises a spatial dimension, due to the urban topology, as well as a temporal one, catching the evolution of the phenomenon over time. The first months as a PhD student should allow to choose between two distinct possible direction. The first one is to perform time series forecasting that exploits the spatial information. The second one, aiming to be more explicative, is to extract spatio-temporal patterns from data.