Density-based community detection in geo-social networks

We propose a density-based model to detect communities of users in geo-social networks that are both socially and spatially cohesive. After formally defining the model and the geo-social distance measure it relies on, we present an algorithm that correctly identifies the underlying communities. We assess the effectiveness of our method using novel quantitative measures on the quality of the discovered communities. We also perform a visual evaluation of the discovered communities, using both real and synthetic datasets. Our results show that the proposed model produces geo-social communities with strong social and spatial cohesiveness, which can not be captured by existing graph or spatial clustering methods.