Clustering in Geo-Social Networks

Abstract

The rapid growth of Geo-Social Networks (GeoSNs) provides a new and rich form of data. Users of GeoSNs can capture their geographic locations and share them with other users via an operation named checkin. Thus, GeoSNs can track the connections (and the time of these connections) of geographic data to their users. In addition, the users are organized in a social network, which can be extended to a heterogeneous network if the connections to places via checkins are also considered. The goal of this paper is to analyze the opportunities in clustering this rich form of data. We first present a model for clustering geographic locations, based on GeoSN data. Then, we discuss how this model can be extended to consider temporal information from checkins. Finally, we study how the accuracy of community detection approaches can be improved by taking into account the checkins of users in a GeoSN.