Opinion Space: Using Dimensionality Reduction to Visualize Public Opinion and Crowdsource Insights
Lecture: Research Exchange | October 6 | 12-1 p.m. | Sutardja Dai Hall, 3rd floor, Banatao Aud.
Ken Goldberg, BCNM, IEOR, EECS, School of Information
Live broadcast at mms://media.citris.berkeley.edu/webcast; Questions can be sent via Yahoo IM to username: citrisevents. The complete schedule for the fall semester is online at http://www.citris-uc.org/events/RE-fall2010. All talks may be viewed on our YouTube channel: http://www.youtube.com/citrisuc.
“Opinion is the medium between ignorance and knowledge.” — Plato
Opinion is notoriously ill-defined and subject to biases and uncertainty principles. On March 15 2010, the U.S. State Department launched Opinion Space, a visualization tool for world opinion developed by students and colleagues at the UC Berkeley Center for New Media. We use dimensionality reduction techniques to display the emerging diversity of viewpoints and collaborative scoring metrics to help the community highlight comments that are most insightful. The display is not based on geography or predetermined categories, but on similarity of opinion. I’ll present the concepts behind the experiment, our experiences working with the State Dept, results so far, and next steps including integrating text analysis using Canonical Correlation Analysis (CCA). You can give it a try at: http://state.gov/opinionspace
Ken Goldberg is Professor of Industrial Engineering and Operations Research at UC Berkeley, with appointments in Electrical Engineering and Computer Science and the School of Information. More information is available at: http://goldberg.berkeley.edu
Available Now: Talk online