Event localization and visualization in social networks

Abstract

Social networks such as Twitter carry important information on ongoing events and as such can be viewed as networks of sensors that monitor and report events in the physical world. An important problem in sensor network literature is that of localization. In the case of monitoring physical events, the localization problem refers to inferring event location from sensor data. In this demonstration, we present a tool that automatically identifies distinct physical events referred to in social network feeds (namely, Twitter feeds) and automatically localizes them. To do so, we designed an algorithm that identifies distinct event signatures in the blogosphere, clusters microblogs based on events they describe, and analyzes the resulting clusters for location information. This information is then translated using the Google Maps API for geo-location, offering a real-time view of ongoing events on a map. To evaluate this tool, we used road traffic related Twitter feeds from San Francisco area in California and corroborate automatic event localization within our service to manually obtained ground truth data. Results show a great correspondence between our automatically determined geo-locations and ground-truth. In the demo, users will be allowed to interact with this and other Twitter data, identify distinct physical events, and locate them in time and space on a map.