FollowMe if you can: a study of mobile crowd sensing with Pokémon go

Mobile crowd sensing becomes a promising solution for massive data collection with the public participation. For the crowd-sharing and crowd-consuming scenarios, the challenges of diversified data sources and quality, as well as the privacy concerns of contributors, make the design and implementation of a sound crowdsourcing platform extremely difficult. In this paper, we present FollowMe, a mobile crowd sensing platform for the popular mobile game Pokémon Go, as a use case to explain the possible design guidelines and solutions to address these issues. It first discusses the incentive mechanisms according to both the quantity and quality of users' contributions. It also applies k-anonymity based solutions to protect contributors' privacy in both scenarios of trustworthy and untrustworthy crowdsourcers. Finally, it proposes a reputation-based filtering solution to detect fake or malicious reports, and a density-based clustering algorithm to find hotspots which can help the prediction of future events. The solutions and discussions in this paper are supposed to be applicable to more complex applications with crowd-sharing and crowd-consuming requirements