Jason Kessler is a lead data scientist at CDK Global, where he analyzes language use and consumer behavior in the online auto-shopping ecosystem. Prior to joining CDK, Jason was the founding data scientist at PlaceIQ and worked as a research scientist for JD Power and Associates. He has published peer-reviewed papers on algorithms and corpora for sentiment and belief analysis, and has sat on program committees and reviewed for several AI and NLP conferences. Most recently, he has conducted research on identifying persuasive and influential language and the visualization of differing corpora.
- Lead (as of Sept. 2016) data scientist at CDK Global, Seattle.
- Founding data scientist at PlaceIQ, NYC.
- Adviser to Votizen, Mountain View, CA. (acquired by Causes)
- Scientist at J.D. Power and Associates, Boulder, CO. Working with Dr. Nicolas Nicolov, I helped to guide the construction of a corpus for structural sentiment analysis and researched ways of automatically annotating structural sentiment relations. Please see our ICWSM 200 and, 2010 papers, as well as our recent Handbook of Linguistic Annotation chapter for details on this effort.
- Summer of 2009:
- Research Intern at Palo Alto Research Center (formally Xerox PARC). I worked on a project in sentiment analysis as a member of the Computing Sciences Lab.
Popular and Trade Press
- Jason S. Kessler. Scattertext: a Browser-Based Tool for Visualizing how Corpora Differ. arXiv. 2017.
- Jason S. Kessler and Nicolas Nicolov. The JDPA Sentiment Corpus for the Automotive Domain. To appear in The Handbook of Linguistic Annotation. Berlin. Nancy Ide and James Pustejovsky, eds., Springer. 2017.
- Jason S. Kessler, Miriam Eckert, Lyndsay Clark and Nicolas Nicolov. The ICWSM 2010 JDPA Sentiment Corpus for the Automotive Domain. 4th International AAAI Conference on Weblogs and Social Media Data Challenge Workshop. Washington, D.C. 2010.
- Jason S. Kessler and Nicolas Nicolov. Targeting Sentiment Expressions through Supervised Ranking of Linguistic Configurations. 3rd International AAAI Conference on Weblogs and Social Media (ICWSM 2009). San Jose, California. 2009.
- Jason S. Kessler. Polling the Blogosphere: a rule-based approach to belief classification. International Conference on Weblogs and Social Media (ICWSM 2008), Seattle, Washington. 2008.
- Theresa Wilson, Paul Hoffmann, Swapna Somasundaran, Jason Kessler, Janyce Wiebe, Yejin Choi, Claire Cardie, Ellen Riloff and Siddharth Patwardhan. OpinionFinder: A System for Subjectivity Analysis. Human Language Technology Conference/Conference on Empirical Methods in Natural Language Processing (HLT-EMNLP 2005). Austin, TX. 2005.
Data and Software
- Language of Closers: Reviews at CDK Global
- Language of Closers: Emails at CDK Global
- Scattertext, a Python term importance and text visualization package.
- You may request the JDPA Sentiment Corpus (used in Kessler and Nicolov  and Kessler et al. ) through the official website.
- The lexicon of terms and multi-word units organized by part-of-speech, veridicality (including facticity) can be found here. These terms, when selected for by syntactic templates outlined in the ICWSM 2008 paper can be used to accurately predict the veridicity of an embedded, finite clause. This an important step in recognizing textual entailment and paraphrase.
- Reviewer, RANLP 2011
- Program committee member, ICWSM 2011
- Session chair, Microblogging, ICWSM 2010
- Reviewer, ACL System Demos 2010
- Program committee member, SocInfo 2010
- Program committee member, IEEE SocCom-2010 Workshop on Finding Synergies Between Texts and Networks, 2010
- Reviewer, Journal of Natural Language Engineering, 2010
- Reviewer, CICLing 2010
- Program committee member, ICWSM 2010
- Reviewer, International Journal of Computers and Applications, 2009
- Reviewer, RANLP 2009
- Program committee member, ICWSM 2009
- External reviewer, MICAI 2008
- External reviewer, CITII 2008
- Ph.D. candidate in Computer Science at Indiana University, Bloomington. My research focused on applying statistical natural language processing techniques for sentiment analysis. Specifically, I explore the compositional way that evaluations is expressed toward discourse entities, a topic my collaborators and I call "structural sentiment."
- In 2005, I received a B.S. in Computer Science from the University of Pittsburgh. While an undergrad, I worked on OpinionFinder, a publicly available system for sentence and expression-level subjectivity analysis.
- A map of Rome showing the location of each Borromini building in the city.
- Richard Serra on Charlie Rose (14 December 2001)
- The Linguist's Search Engine.
- The Gallery of "Misused" Quotation Marks
- SIL's glossary of linguistic terms.
- My Erdös number is less than or equal to 5 (via Claire Cardie ~ Raymond J. Mooney ~ Wolfgang Maass ~ Andras Hajnal)
- A slightly outdated guide by John "Verm" Sherman to the Hueco Scale for grading boulder problems.