Computational Analysis of Text

This project applies Natural Language Processing Techniques on a webscraped corpus of government documents to identify when the U.S. /Congress and President/ discussed or imposed economic sanctions.  We rely on supervised machine-learning techniques to extract key-word combinations related to the economic pressure imposed by the U.S. and the associated violations. Violations can be issues related to terrorism, narcotics-control, nuclear weapons to issues relevant to democratization and human rights, and trade practices. We also cover a range of varying economic foreign policy tools imposed on violating countries.

Team: Ashrakat Elshehawy, Nikolay Marinov and Federico Nanni of Uni Mannheim and Jordan Tama of GWU. 

This is research in progress.  A completed example of this research (based on a dictionary approach) is ``Quantifying Attention to Foreign Elections with Text Analysis of U.S. Congress and the Presidency,'' by Ashrakat Elshehawy, Nikolay Marinov and Federico Nanni of the University of Mannheim derive a new measure of attention to foreign elections.  The authors thank seminar participants at the University of New Hampshire, February 2017, for feedback and comments on this and earlier versions.  Marinov thanks the Electoral Integrity Project, Sydney, where some of the early work was done, March 2016.  

The research note is available from SSRN

One substantively-focused project with the new data is Varieties of Unilateral Enforcement of International Human Rights Standards (with Daniela Donno, Ashrakat Elshehawy) [Link]