Privacy Leakage in Event-based Social Networks: A Meetup Case Study

Abstract

Event-based social networks (EBSNs) are increasingly popular since they provide platforms on which online and offline activities are combined. Despite the increasing interest in EBSNs, little research has paid attention to the privacy issues coming from the unique features of EBSNs; the on-site information of users is highly relevant to real lives. In this paper, we try to investigate privacy leakages in Meetup, one of the most popular EBSN service. More specifically, we answer what private information can be inferred from the site's publicly available data. To this end, we conduct a measurement study by crawling webpages from Meetup containing 240K groups, 8.9M users, 27M group affiliations and 78M topical interests. By analyzing the dataset, we find that LGBT status of users, which is one of the most sensitive privacy information, can be predicted with 93% accuracy. Finally we discuss the cause of the privacy leakage on EBSNs and its possible ensuing damages.