[Colloquium] REMINDER: Matt Taddy | TOMORROW, Wednesday, January 29, 2014

[Olivia Lin]

*Please forward this announcement to your colleagues who may be
interested. *



*Visualizing and Measuring Social Data*

*Matt Taddy*

Associate Professor, Econometrics and Statistics, University of Chicago
Booth School

*TOMORROW*

*Wednesday, January 29, 2014, 3:00pm-4:30 pm*

*Kathleen A. Zar Room, John Crerar Library*

*Cookies and Refreshments will be served*

*BIOGRAPHY*

Matt Taddy is an associate professor of econometrics and Statistics and
Neubauer Family Faculty Fellow at the University of Chicago Booth School of
Business. His research is focused on statistical methodology and data
mining, driven by applications in business and engineering. He developed
and teaches the MBA 'Data Mining' course at Chicago Booth.

Taddy works on building robust solutions for large scale data analysis
problems. This involves dimension reduction techniques for massive datasets
and development of models for inference on the output of these algorithms.
Applications are ongoing in consumer database mining, digital marketing,
analysis and optimization of computer simulators, and in text mining for
analysis of social media, financial news, and political speech. He has
collaborated both with small start-ups and with large research agencies
including NASA Ames, Lawrence Livermore, Sandia, and Los Alamos National
Laboratories.

Taddy earned his PhD in applied Math and Statistics in 2008 from the
University of California, Santa Cruz, as well as a BA in Philosophy and
Mathematics and in Mathematical Statistics from McGill University. He
joined the University of Chicago Booth faculty in 2008
*ABSTRACT*

This presentation will explore 'Big Data' problems: inference from
unstructured data that is too large to analyze, or even store, on a single
computer. In social science, this will typically involve some amount of
text analysis, even if just to extract variables of interest from the
original data. An example application could involve the internet browsing
history for a number of individuals, their physical location, a corpus of
text associated with these individuals (both from the web pages they visit,
and text generated on social media), and, say, purchases by these
individuals. Or, we may wish to understand the relationship between a
number of economic or political variables and the co-movement of topics,
terms, and tone in the news and on social media.

In such text mining applications there is an interplay between two
visualization strategies: plotting predictions and factors that summarize
the information you have mined from the text, and looking at the role
played by individual phrases or words in the model driving this
summarization. Getting these two modes of visualization to work together is
key to communicating and understanding results. In this talk,
Dr. Taddy will cover the basic idea of how the statistical models behind
the analysis work, and use this to understand what one might want to be
plotting. The main goal is then to illustrate how the two modes of
visualization work together. This will be shown through
example applications including financial news, political speech, and social
media.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20140128/30e8a228/attachment.htm