Citation

Bahir E, Peled A. Soc. Sci. Comput. Rev. 2013; 31(4): 458-470.

Copyright

(Copyright © 2013, Duke University Press)

DOI

10.1177/0894439313483689

PMID

unavailable

Abstract

In recent years, several technological advancements have changed the lives of millions throughout the globe. These include broadband Internet wireless access, advanced mobile platforms and smartphones including accurate global positioning system capabilities, and the introduction of social networks. The fusion of these technological advances led to the massive adoption of mobile platform-operated social networking applications and unleashed new real-time and on-site social information. The ability to generate content anywhere and anytime leads to a detectable projection of real-life events on geo-social networks (GSN). For example, in preparation for a rally, the geo-social activity may precede the actual event, allowing predictive capabilities. Alternatively, in a natural event such as a wildfire, early content generated in the proximity of the event may allow early identification of the event and the assessment of its physical boundaries. In this article, we propose to use the massive and rapidly accumulating information communicated within GSN to identify and track major events and present a proof of concept. We discuss means and methods to retrieve relevant information from the networks, through a set of adequate spatial, temporal, and textual filters. Our preliminary empirical results corroborate our assumptions and show that major events may have detectable "abnormal" impact on GSN activities, which allows prompt identification and real-time tracking. Our approach is expected to pave the way to the development of real-time systems and algorithms for early identification and geographical tracking of major events.

Language: en