Automatic mapping and modeling of human networks

Mobile telephones, company ID badges, and similar common devices form a sensor network which can be used to map human activity, and especially human interactions. The most informative sensor data seem to be measurements of person-to-person proximity, and statistics of vocalization and body movement measurements. Using this data to model individual behavior as a stochastic process allows prediction of future activity, with the greatest predictive power obtained by modeling the interactions between individual processes. Experiments show that between 40% and 95% of the variance in human behavior may be explained by such models.