Analisi Spazio-Temporale di Messaggi Twitter per l'Identificazione di Eventi

Density based clustering is proposed as an effective way to perform geographic and temporal exploration of messages freely generated within social contexts, in order to reveal and map their latent spatio-temporal structure. The approach is exemplified to identify geographic regions where many geotagged Twitter messages about a given event have been created, possibly in the same time period in the case of aperiodic event, or at regular timestamps in the case of periodic events.