Footprints

by Columbia University

We will attempt to provide geo-location privacy through decoy queries

We will attempt to create a middleware between the Android OS and various applications, so as to provide fake decoy geolocation information to third party services. For instance, whenever a user leaves their workplace to go to a
restaurant, decoy profiles of that user will be visiting other locations in the same city. If the user uses an app to search for restaurants near her current location, our framework will issue similar searches in the background from
fake locations. In this sense, for every real account and location, a set of decoy accounts and requests will be presented to the service. This way, users will be able to enjoy the benefits of location-sensitive context without
giving away a long-term profile of their behavior

We will use OpenPath data as a pool of real user traces. Our decoy accounts will be following real routes in NYC based on OpenPath data.

No private information will be disclosed to any third party. All data will be anonymized and stored securely. We are only interested in the GPS traces and not the types of devices or the names of the participants

Currently this is a course project but our goal is to integrate it with Android's Privacy
Guard and provide a generic tool for geo-location privacy. All code will be open source once the project is finished. If we obtain encouraging results, we will pursue a publication.