Analysis of Geolocalized Social Networks Based on Simplicial Complexes


A common issue in network analysis consists in the detection and characterization of the key vertices and communities. To this purpose, visualization tools could be of great help to support domain experts in analyzing this kind of data. However, the size of real networks can seriously affect the practical usage of these tools, thus, requiring the definition of suitable simplification procedures that preserve the core network information. In this work, we focus on geolocalized social networks, and we describe a tool for the analysis of this kind of data based on topological information. Supported by recent trends in network analysis, our approach uses simplicial complexes as a model for social networks. A homology-preserving simplification is used for dealing with the data complexity and for reducing the information to be visualized to the essential. By combining the representation based on simplicial complexes and the simplification tool, we can efficiently retrieve topological information useful for the network analysis. Both the effectiveness and scalability of our approach are experimentally demonstrated.

Proceedings of the 9th ACM SIGSPATIAL Workshop on Location-based Social Networks