Exploring Real Mobility Data with M-Atlas R. Trasarti, S. Rinzivillo, F. Pinelli, M. Nanni, A. Monreale, C. Renso, D. Pedreschi, and F. Giannotti Pisa KDD Laboratory, ISTI - CNR, Italy 1 Introduction Research on moving-object data analysis has been recently fostered by the widespread diffusion of new techniques and systems for monitoring, collecting and storing loca- tion aware data, generated by a wealth of technological infrastructures, such as GPS positioning and wireless networks. These have made available massive repositories of spatio-temporal data recording human mobile activities, that call for suitable analytical methods, capable of enabling the development of innovative, location-aware applica- tions [3]. The M-Atlas is the evolution of the system presented in [5] allows to handle the whole knowledge discovery process from mobility data. The analysis capabilities of M-Atlas system have been applied onto a massive real life GPS dataset, obtained from 17,000 vehicles with on-board GPS receivers under a specific car insurance contract, tracked during one week of ordinary mobile activity in the urban area of the city of Milan; the dataset contains more than 2 million observations leading to a set of more than 200,000 trajectories (see Fig.1). 2 The M-Atlas System A system able to master the complexity of the knowledge discovery process over mo- bility data needs to support at least four functionalities: (i) trajectory data need to be created, stored and queried through spatio temporal primitives; (ii) trajectory models and patterns representing collective behavior have to be extracted using trajectory min- ing algorithms; (iii) such patterns and models have to be represented and stored in order to be re-used or combined; (iv) new mining algorithms may be added. The M-Atlas system allows the user to combine all these aspects through an innovative Data Mining Query Language (DMQL). This language can be used to express the whole knowledge discovery process as a sequence of queries to be submitted to the system. The GUI in- terface gives the user the possibility to use pre-defined analysis (i.e. O/D Matrix) or to use the console to write down his/her own DMQL queries. In the next sections we will give a short example of the capabilities of the system on the real dataset of Milan described above. 3 Understanding Mobility To grasp a general vision on the dataset we performed a series of statistical analysis on the dataset (the charts representing the results are shown in Figure 2): J.L. Balc´ azar et al. (Eds.): ECML PKDD 2010, Part III, LNAI 6323, pp. 624–627, 2010. c  Springer-Verlag Berlin Heidelberg 2010