ENBIS-11 in Coimbra

4 – 8 September 2011 Abstract submission: 1 January – 25 June 2011

Collective Prediction in Digital Social Networks Through Discovery of Collaborative Innovation Networks

5 September 2011, 15:15 – 16:15


Submitted by
Peter Gloor
Peter Gloor
MIT Sloan School of Management
The emergence of online social networks opens up unprecedented opportunities to read the collective mind, discovering emergent trends while they are still being hatched by small groups of creative individuals. The Web has become a mirror of the real world, allowing researchers in predictive analytics to study and better understand why some new ideas change our lives, while others never make it from the drawing board of the innovator. This talk introduces a wide range of projects doing analysis of large corpora of digital traces of human activity, in particular the Web, Blogs, online forums, social networking sites, e-mail archives, phone logs, and face-to-face interaction through sociometric badges. In our research we have identified interaction patterns of successful teams and organizations through dynamic social network analysis (DSNA). On the global level, we correlate discussion about movies on IMDB with Academy Awards and box office returns, and Blog buzz with the outcome of political elections. On the organizational level we compare performance metrics such as revenue, productivity, peer ratings, and customer satisfaction with social network metrics in a variety of settings, including, for example, Eclipse open source programmers. One of the main outputs of our work is the Condor software tool, formerly known as TeCFlow, which generates dynamic visualizations of social networks by mining communication archives such as e-mail, phone logs, and blogs. In a series of projects we collaborate with the MIT Media Lab, which has developed sociometric badges that can measure the location and behavioral characteristics—such as gestures and tone of voice—of their wearers. These sociometric badges have proven to be a powerful tool that can connect the impact of behaviors such as speech patterns on business outcomes such as the success of marketing campaigns, or the speed of recovery of hospital patients, predicting performance of teams and individuals as well as their personality characteristics.

Return to programme