Cloud Seminar: Learning About the World Through Social Media

Cloud Seminar: Learning About the World Through Social Media Seminar | February 13 | 11 a.m.-12 p.m. | 380 Soda Hall
Speaker/Performer: Emre Kiciman, Microsoft Research Sponsor: Electrical Engineering and Computer Sciences (EECS)
Social media, namely Twitter and Facebook, provides a finer-grained,
larger-scale digital record of human interactions than has been
available before. Unfortunately, social media is also a noisy,
ambiguous and biased medium. In this talk, I will describe two
projects that begin to untangle this messiness. The first is an entity
recognition system that analyzes the Twitter firehose, adding
structure to otherwise unstructured text. The second is a large-scale
study of self-reporting bias that uses ground-truth data to analyze
the factors that influence people to tweet about real-world events. I
will touch on two very different applications of such social media
analysis: how extracting information about peoples real-world
activities can help us answer search queries; and a study of social
media usage in the context of the Mexico drug war.

Biography:
Emre Kiciman is a researcher in the Internet Services Research Center
at Microsoft Research. His interests are in social network and media
analysis, particularly as they relate to social search. His current
projects focus on mining new kinds of information about the world from
social media, and using social network data for search ranking. Emres
previous research interests include JavaScript application monitoring
and optimization, as well as the reliability of Internet services
architectures and operations. Emre earned a Ph.D. and M.S. in computer
science from Stanford University, and a B.S. in Electrical Engineering
and Computer Science from U.C. Berkeley.
Event Contact: alig@cs.berkeley.edu