A privacy-preserving personal sensor data ecosystem

Abstract

Despite the ubiquity of passively-collected sensor data (primarily attained via smartphones), there does not currently exist a comprehensive system for authorizing the collection of such data, collecting, storing, analyzing, and visualizing it in a manner that preserves the privacy of the user generating the data. This thesis shows the design and implementation of such a system, named openPDS, from both the client and server perspectives. Two server-side components are implemented: a centralized registry server for authentication and authorization of all entities in the system, and a distributed Personal Data Store that allows analysis to be run against the stored sensor data and aggregated across multiple Personal Data Stores in a privacy-preserving fashion. The client, implemented for the Android mobile phone operating system, makes use of the Funf Open Sensing framework to collect data and adds the ability for users to authenticate against the registry server, authorize third-party applications to analyze data once it reaches their Personal Data Store, and finally, visualize the result of such analysis within a mobile phone or web browser. A number of example quantified-self and social applications are built on top of this framework to demonstrate feasibility of the system from both development and user perspectives.