Decoding Collective Wisdom in Social Networks

When: Wednesday, May 02, 2012 at 2:00 pm
Where: DA 5th fl
Speaker: Professor Brian Uzzi
Organization: Northwestern University, Evanston, IL, Co-Director, Northwestern Inst. on Complexity Systs. and Network Science (NICO), Richard L. Thomas Distinguished Professor of Leadership and Organizational Change, Kellogg School of Management, Professor of Sociology, Weinberg College of Arts & Sciences, Professor of Industrial Engineering and Management Sciences, McCormick Engineering
Sponsor: CCNR and Lazar Lab Seminar

Although more than 100 years have passed since Sir Francis Galton observed that “wisdom of the crowd” can outperform experts, methods to identify and exploit collective wisdom are now emerging. We develop three methods for finding collective wisdom in e-communication network data. First, we demonstrate how the content of e-communication data signals an actor’s activation level, an emotional state that predicts when, whether and how much behavior will follow, even as the actors themselves may be oblivious to the link between their word choice and emotional state signaling. Second, we observe how the collective’s network of words used in e-communications can portend a system’s future volatility better than other measures. Third, we observe how synchronicity in behavior is linked to performance and distinctive communication patterns. Our tests are made on high resolution data covering the second by second securities trades of day traders and the more than 3 million IMs sent and received among them and their contacts.