Résumé en anglais

Traffic data coming from sensor networks have prompted a wide range of research issues related with Transportation Information Systems. These data are usually represented by large and complex spatio-temporal series. This paper presents a new approach to manage rough data coming from static georeferenced sensors. Our work is based on combining analytic methods to process sensor data and proposing an architecture for an information system dedicated to road traffic. It is being conducted within a project which uses real data, generated by 1000 sensors, during 3 years, in a french big city.