Although the representation of networks as real-world systems or processes is a well-known topic that has been extensively discussed in the literature, the impact of geographic distributions of actors (network nodes) has not been fully investigated. The geographic configuration of networks could have a significant relationship with the logic representation, affecting many common operations on networks, such as community detection.

Liu and Huang propose a community detection method that takes into account these geographic distributions in order to improve the accuracy of results according to common metrics, under the realistic assumption that some networks are more influenced by location than others. The authors have performed extensive experiments on synthetic and real-world datasets that have showed interesting results in terms of performance.

I liked this paper for its approach, focus, and concise discussion of results. As with most works that relate real-word parameters (geographic location in this case) to logic representations, it could inspire many other studies, both in this context and beyond.