Fostering the expressive power of individual mobility networks for fleet trajectories modeling

The quick evolution and wide diffusion of technologies for the localization of devices, especially smartphones and vehicles' GPS, is leading to the production and collection of large and diversified traces of human mobility. This large availability of mobility data allows us to investigate complex phenomena about human movement and to study the human behavior. However this abundance of raw data usually comes with few additional information about the points collected. Hence, in order to unlock this potential, we need to define methods for processing and analyzing mobility data. In this thesis we foster the expressive power of Individual Mobility Networks (IMNs), a data model describing a user mobility, to create a procedure to annotate the locations where the users have stopped. We have called the combination of IMNs with these labels Annotated IMNs (AIMNs). They allow a generalization which makes the locations and the vehicles comparable. The procedure exploits a set of features based on different characteristics of a location. Then, by applying a clustering process, obtains a small set of labels that can be used to classify the vehicles according to the type of locations they visit. We tested the algorithm on a dataset of trucks moving in Greece. The results show that the AIMNs can enable detailed analysis of urban areas and the planning for advanced mobility applications.