Section 1

research

Cline Center Phoenix Event Data Research

 

As a member of the Open Event Data Alliance (OEDA), the Cline Center is developing purely algorithmic methods for extracting events from news media. The Cline Center’s systems are based on the Python Engine for Text Resolution And Related Coding Hierarchy (PETRARCH-2). PETRARCH is a third-generation system built on more than 25 years of engineering and research at the Kansas Event Data System (KEDS) and is similar to the US government’s Integrated Crisis Early Warning System (ICEWS). Our pipeline processed more than 14 million documents to produce the Cline Center Historical Phoenix Event Data, which you can download here.

Work on machine-generated event data at the Cline Center leverages the Cline Center’s Global News Archive and text analytic methods to document the agents, locations, and issues at stake in a wide variety of conflict, cooperation and communicative events. Cline Center-supported researchers from a variety of disciplines are working to improve geo-location, topic identification, overall accuracy, and speed of event data extraction algorithms, and are also developing a user interface for customized PETRARCH applications.

We are also developing hybrid event data production systems that combine human judgement and machine efficiency in our SPEED data project.