This document presents an information extraction system that uses indoor location and activity data to summarize what sessions participants are attending at a conference venue. The system architecture includes sensors to collect physical location data, agents to reason about the context data, and an ontology to model context information. It evaluates tracking 268 of 736 participants over 3 days at the conference, collecting 2.6 million location records. By analyzing participant location clusters for each session time slot, the system aims to detect which sessions a participant spent time in and predict their next session. Future work involves improving prediction capabilities and applying the system to elder care environments.