New Directions in Analyzing Text as Data

CONFERENCE:   

Sixth Annual New Directions in Analyzing Text as Data Conference.

WHEN:
Friday, October 16, 2015, 8:30 am to
Saturday, October 17, 2015, 5:00 pm
WHERE:
NYU School of Law Event Space,
108 West Third Street
New York University, New York, NY.
WHAT:
Paper presenters have been selected and notified.  Thank you to all those who submitted.

Please feel free to reach out to Rebecca Liebe [rebecca.liebe@nyu.edu] with any questions.

CONFERENCE INFORMATION:

The main purpose of this conference is to bring together researchers from social science, computer science and linguistics to investigate new approaches to utilizing text in (social) science research. Text has always been a valuable resource for research. However, recent developments in digital archiving as well as breakthroughs in automatic language-processing methodologies from the fields of information retrieval, natural language processing, and machine learning are creating unprecedented opportunities for searching, categorizing, and extracting social science information from text.

Previous conferences took place at Harvard University, Northwestern University and London School of Economics. Selection of participants and papers for the conferences is the responsibility of a team led by Arthur Spirling (Harvard/NYU) and Ken Benoit (LSE) and consists of Noah Smith (University of Washington), Justin Grimmer (Stanford) and Amber Boydstun (UC Davis)

This two-day invitation-only conference draws together scholars from many different universities and disciplines to discuss developments in text as data research.  We anticipate that our conference budget will cover the airfare and hotel expenses of paper presenters and discussants, and hope that some funds will also be available to defray the costs for other attendees (see Conference Expenses for specific details).

Sponsored by:

Cube_Logo_NoLights_384x216

Registration

Registration for this event is now closed.

Attendance is by invitation only and we have met the capacity for this event.

If you would like information on future conferences, please contact Joshua Choe at this email.

Sessions Information - Friday, October 16th

Friday, October 16, 2015

Check-in, Breakfast & Welcome

7:15am to 8:15am – Check-in and Continental Breakfast (Lipton Hall Foyer)
8:15am to 8:30am – Welcome from the Conference Chair (Lipton Hall)

 

Session 1: Give a Dog a (funny) Bone

(Note:  All sessions will be in Lipton Hall)
8:30am to 10:00am

1a) Who’s Laughing Now?  Applying Text Analysis to Humor in Federal Reserve Meetings
Speaker: Amber Boydstun
Click here to access speaker paper (password required)

1b) The dog whistle detector: searching for lexical evidence of coded partisan speech
Speaker: David Mimno
Click here to access speaker paper (password required)

Discussants: John Patty and Mona Vakilifathi

 

10:00am to 10:15am  – Break (Lipton Hall Foyer)

 

Session 2: Master of All We Survey

10:15am to 11:45am

2a) Automated Coding of Open Ended Survey Responses
Speaker: Dallas Card
Click here to access speaker paper (password required)

2b) Predicting Traditional Surveys From Social Media With Topic Models
Speaker: Mark Dredze
Click here to access speaker paper (password required)

Discussants: Ben Lauderdale and Reagan Rose

 

11:45am to 1:00pm – Lunch Faculty Club – Lipton Hall

 

Session 3: History, Repeating

1:00pm to 2:30pm

3a) Who, What, When, Where, and Why?  A Computational Approach to Understanding Historical Events Using State Department Cables
Speaker: Allison Chaney
Click here to access speaker paper (password required)

3b) Comparing Methods for Generating Large Scale Political Event Data Sets
Speaker: Phillip Schrodt
Click here to access speaker paper (password required)

Discussant: Brandon Stewart

 

2:30pm to 2:45pm  – Break (Lipton Hall Foyer)

 

Session 4: Health and Safety

2:45pm to 4:15pm

4a) Climbing Mount Obamacare:  Experimentally Optimized Textual Treatments 
Speaker: Nick Beauchamp
Click here to access speaker paper (password required)

4b) Conducting sparse feature selection on arbitrarily long phrases in text corpora with a focus on interpretability
Speaker: Luke Miratrix
Click here to access speaker paper (password required)

Discussant: Marc Ratkovic

 

4:15pm to 4:30pm  – Break (Lipton Hall Foyer)

 

Session 5: Being Judgmental

4:30pm to 6:00pm

5a) Using Opinion Language to Predict Review in a Judicial Hierarchy
Speaker: Matthew Hitt
Click here to access speaker paper (password required)

5b) Reading Between the Emails:  Understanding Gendered Patterns of Email Communication in Local Government Organizations
Speaker: Matthew Denny
Click here to access speaker paper (password required)

Discussants: Paul Kellstedt and Ashley Anderson

Sessions Information - Saturday, October 17th

Saturday, October 17, 2015

8:00am to 9:00am – Breakfast

 

Session 6: News and Views

9:00am to 10:30am

6a) Coding the tone of political texts: Do bad things happen to good people?
Speaker: Maurits van der Veen
Click here to access speaker paper (password required)

6b) Sentiment in Newsprint and its Economic Effects:  The good, the Bad, and the Uncertain
Speakers: Michelle Alexopoulos and Betsy Barry
Click here to access speaker paper (password required)

Discussant: Christopher Zorn

 

10:30am to 10:45am  – Break (Lipton Hall Foyer)

 

Session 7: Dimensions in Space and Time

10:45am to 12:15pm

7a) Matching Methods for High-Dimensional Data With Applications to Text
Speaker: Molly Roberts
Click here to access speaker paper (password required)

7b) Topic-Sentiment Model with Document-Level Covariates
Speaker: Christine Kuang
Click here to access speaker paper (password required)

Discussant: Will Lowe

 

12:15pm to 1:30pm – Lunch (Faculty Club-Lipton Hall)

 

Session 8: Rhetorical Device(s)

1:30am to 3:00pm

8a) Measures of Ideology: Agendas, and Positions on Agendas
Speaker: Gaurev Sood
Click here to access speaker paper (password required)

8b) Rhetorical Patterns in Legislative Speech
Speaker: Jacob Eisenstein
Click here to access speaker paper (password required)

Discussant: David Sontag

 

3:00pm to 3:15pm  – Break (Lipton Hall Foyer)

 

Session 9: Faith, Hope and Clarity

3:15pm to 4:45pm

9a) More Than Unigrams Can Say 
Speaker:  Paul Nulty
Click here to access speaker paper (password required)

9b) The Unreliability of Measures of Intercoder Reliability, and What to do About It
Speaker: Justin Grimmer
Click here to access speaker paper (password required)

Discussant: David Bamman

 

4:45 to 5:00pm – Closing Remarks

New Directions in Text as Data

Lipton Hall & Faculty Club
(also known as D’Agostino Hall – NYU School of Law Event Space)
108 West 3rd Street, NY, NY

Get Directions

About the NYU Center for Data Science

Established in 2012, NYU’s Center for Data Science (CDS) seeks to develop knowledge in fields as diverse as astronomy, genomics, medicine, physics, and political science. CDS acts as a hub for interdisciplinary research on related to data science. It focuses on fostering a network of researchers in the natural and social sciences, as well as the humanities. With the goal of training data scientists of the future, CDS recently introduced a Master of Science in Data Science degree program and will soon launch a PhD program. NYU’s pioneering efforts in data science continue the university’s long tradition of innovation. Since Samuel Morse, inventor of the electric telegraph, joined the faculty in 1831, NYU has nurtured numerous inventors, scientists, and entrepreneurs. In the last century alone, NYU alumni and entrepreneurs have established companies that transformed industries such as ADP, Def Jam, Bloomberg, Southwest Airlines, Etsy, Pinterest, and Twitter.

Register