Social discovery using longitudinal spatial smartphone sensing

A new generation of social discovery and crowd-sourced location-based services promise to integrate mobile computing into our lives more than ever. An essential component in these services is the neighbor or peer selection, that is also needed for peer-to-peer (P2P) mobile applications and multiplayer mobile gaming. One major challenge is posed by the interpretation of the information collected by the smartphones (about nearby devices, locations) to bring the user closer to context-awareness and informed peer selection. We introduce a framework and a corresponding application (ConnectEnc) for mobile peer rating using a multi-dimensional metric scheme based on encounter and location sensing. ConnectEnc maintains a history of discovered nearby devices and locations, and rates them using an array of metrics ranging in complexity from simple (encounter frequency) to complex and novel (location vector and matrix similarity). We have developed and deployed ConnectEnc application on Android and Nokia N810 platform to measure the link between the scores of proposed filters and device selection by the user. Results from the deployment shows that statistically high correlation exists between ConnectEnc recommendations and user selections. Our framework is distributed, modular and can serve as a peer selection platform for other mobile protocols based on specific requirements. It runs on individual devices and does not require data sharing or interaction; and hence can bootstrap other recommendation, cooperation or reputation systems.