Abstract

Abstract. Nowadays, people conduct a lot of activities with their online social networks. With the large amount of social data available, quanti-tative analysis of social influence becomes feasible. In this PhD project, we aim to study users ’ social influence at the community level, mainly because users in social networks are naturally organized in communities and communities play fundamental roles in understanding social behav-iors and social phenomenons. Through experiments with a location-based social networks dataset, we start by demonstrating communities ’ influ-ence on users ’ mobility, and then we focus on the influence of leaders in the communities. As a next step, we intend to detect users that act as structural hole spanners and analyze their social influence across different communities. Based on these studies, we plan to propose a unified ap-proach to quantify users ’ social influence and investigate its applications, for example, in social interaction and behavior analysis. 1