Personal information gathered via sensors carried by humans (in cars or on mobile phones), and personal information stored on community web sites, if merged, can open new avenues for research. For example, the information about a user's friend-network (as available from the Orkut website) can be used to restrict access to information obtained via sensors. At the same time, information obtained via sensors can be added to the information stored on the web (e.g images, video, location etc), tagged and queried on demand. In this project, I plan to discover ways of symbiotically using these two hitherto disjoint sets of personal information. The goal is to build an infrastructure for sensing, collecting, merging, analyzing and using personal information in creating novel and secure usage models.
I have taken the first step in this direction. I have a system that logs personal information (such as location, wireless devices in the proximity, incoming/outgoing calls, incoming/outgoing SMSs) on the mobile phone of the user. I have developed algorithms that use this information for predicting future locations and activities of the user. I'm currently building an infrastructure for mining and analyzing data from community websites.