Social ties and checkin sites: connections and latent structures in location-based social networks

Location-based social networks integrate location-based facilities with social connectivity for delivering a variety of services, enhancing user experience, emergency/disaster management, and streamlining business processes. A number of recent research efforts have studied relationships between geolocation and social connectivity, social connectivity and preferences, and node attributes and strength of social ties. These efforts have successfully demonstrated prediction of various attributes based on social connectivity, mobility, dynamic checkin information, etc., including prediction of user location as well as future checkin locations. In this paper, we study the relationship between shared checkin locations and the structure and nature of social ties. We argue that typical LSBNs are in fact composed of layers of networks of varying structure and function and that it is possible to deconcolve these networks through effective statistical analysis of shared checkins. In this context, we pose and validate the following hypotheses: (1) A large number of shared checkins imply social connectivity; however, social connectivity does not imply statistically large number of shared checkins; (2) entities in social ties that share a large number of checkins tend to be strongly clustered. We hypothesize that such strong ties (e.g., family ties and friendships) carry higher influence compared to weaker ties (mere acquaintances) in the social network; (3) social ties that have statistically fewer shared checkins (weak ties) tend to be less clustered than the underlying (baseline) network. We hypothesize that such ties (e.g., professional ties, friends of friends, and acquaintances) carry less influence; and (4) social ties that have statistically large number of shared checkins (strong ties) tend to be relatively more dynamic when compared to weak ties over a period of time. We present statistical models and validate our hypotheses on real datasets. Our conclusions can significantly enhance flow of information and influence in the network by suitably leveraging the distinct relationships captured in the deconvolved networks.