We aim at devising the basic principles of a data-driven science for social mining, i.e., the conceptual and computational tools to create abstract models of social behavior and dynamics based on novel analytical paradigm at the crossroad of network science and data mining from big data of human activities. We will concentrate on three aspects of social behavior: the patterns of human mobility, the diffusion dynamics in social networks, and the patterns of success in sports.
The behavioral models we aim at should represent profiles of human dynamics at different scales, replicable, re-applicable and re-localizable to different context and different geographies. Such models are expected to shed a deeper light on the way cities work, as well as on the diffusion phenomena in social networks and the understanding of how success is reached in sport, based on novel network representations of human dynamics and novel social sensing and mining techniques for capturing and analysing the digital traces of human activities. We shall adopt a data-driven science approach, where theory emerges from large-scale experiments, based on the unique experiences and datasets made available by our KDD Lab. at Univ. Pisa, in collaboration with the Center of Network Science at CEU. We shall build on our vast experience in analysing large-scale mobility data, such as GPS tracks, GSM (mobile phone records) datasets and many more big data sources, which will enable to study the evolutionary dynamics of diffusion, success and movement phenomena.