"Will check-in for badges": Understanding bias and misbehavior on location-based social networks

Abstract

Social computing researchers are using data from locationbased social networks (LBSN), e.g., "Check-in" traces, as approximations of human movement. Recent work has questioned the validity of this approach, showing large discrepancies between check-in data and actual user mobility. To further validate and understand such discrepancies, we perform a crowdsourced study of Foursquare users that seeks to a) quantify bias and misrepresentation in check-in datasets and the impact of self-selection in prior studies, and b) understand the motivations behind misrepresentation of check-ins, and the potential impact of any system changes designed to curtail such misbehavior. Our results confirm the presence of significant misrepresentation of location check-ins on Foursquare. They also show that while "extraneous" check-ins are motivated by external rewards provided by the system, "missing" check-ins are motivated by personal concerns such as location privacy. Finally, we discuss the broader implications of our findings to the use of check-in datasets in future research on human mobility.