By Peixoto, J.; Moreira, A.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Capturing, representing, modelling and visualizing the dynamics of urban mobility have been attracting the interest of the research community recently. One of the drivers for recent work in this area is the availability of large datasets representing many aspects of the urban dynamics. Applications for these studies are diverse and include urban planning, security, intelligent transportation systems and many others. Quite often, the proposed approaches are highly dependent on the data type. This paper describes the definition of a set of basic concepts for the representation and processing of spatio-temporal data, sufficiently flexible to deal with various types of mobility data and to support multiple forms of processing and visualization of the urban mobility. A place learning algorithm is also described to illustrate the flexibility of the proposed framework. Available results obtained by the integration of geometric and symbolic data reveal the adequacy of the proposed concepts, and uncover new possibilities for the fusion of heterogeneous datasets.