• big data
• data mining
• mobilità

In the era of Big Data a huge amount of information are available from every sin- gle citizen of our hyper-connected world. A simple smartphone can collect data with different kinds of information: a big part of these are related to mobility. A smartphone is connected to networks, such as GSM, GPS, Internet (and then social networks): each of them can provide us information about where, how and why the user is moving across space and time. Data integration has a key role in this understanding process: the combination of different data sources increases the value of the extracted knowledge, even though such integration task is often not trivial. This thesis aim to represent a step toward a reliable Mobility Analysis framework, capable to exploit the richness of the spatio-temporal data nowadays available. The work done is an exploration of meaningful open challenges, from an efficient Map Matching of low sampling GPS data to Inferring Human Activities from GPS tracks. A further experimentation has been performed over GSM and Twitter data, in order to detect and recognize significant events in terms of people presence and related tweets. Another promising perspective is the use of such extracted knowledge to enrich actual geospatial Datasets with a ’Wisdom of the crowd’ dimension to derive, for instance, routing policies over road networks: most chosen paths among usual drivers are more meaningful than simple shortest paths.