Quantifying Information Flow in Meetings
We integrate speech, screen-share, and note logs to quantify how information flows—who influences whom and how topics evolve—revealing bottlenecks, imbalances, and convergence with objective metrics.
This project establishes a measurement framework for the flow of information and the formation of shared understanding in meetings. We synchronize turn-taking, keyword transitions, references to artifacts (slides, documents, notes), and decision events to compute: (1) participant-level network metrics (in/out centrality, reply ratio, response-latency distributions), (2) topic dynamics metrics (Discussion Expansion, Local Intensity, Topic Variety), and (3) blockage/gap indicators (unanswered-question rate, isolated topics, unresolved actions). Facilitation moves (e.g., reframing, summarizing, targeted turn allocation) are time-stamped, enabling causal modeling of pre/post changes in these metrics to identify which intervention types and timings improve information flow. Deliverables include a real-time and after-action dashboard, standardized metric definitions with analysis scripts, and a practice guide for bias detection, participation rebalancing, and convergence support—providing an evidence base for improving meetings in education, research, and industry.