Brown Bag Series
Automated Event Data Collection: Extending the BFRS Political Violence Dataset
Friday, November 22, 2019
A-16, Academic Block, LUMS
Work in collaboration with Fareed Zafar and Hina Usman
We demonstrate the viability of automated data collection about the occurrence of specific events reported in newspapers. Specifically, the existing BFRS dataset codes political violence from 1988 to 2011, as captured by human researchers manually reading and extracting information from the Lahore edition of Dawn newspaper. We create a similar dataset from 2010 to the present day by automating categorization of events using textual analysis with pattern recognition. Automation improves the speed of data collection and reduces marginal cost to near-zero. We evaluate the quality of our machine-collected data, report political violence trends in the nine years since BFRS was collected, and discuss the viability of deploying these methods to other areas of work.
Dr. Ali Hasanain
Dr. Ali Hasanain is Assistant Professor of Economics at the Lahore University of Management Sciences. He is also a member of EGAP, a CDPR fellow, and a faculty advisor at the Technology for People Initiative. Moreover, 2014 to 2016, he was a Global Leaders Fellow at Oxford and Princeton universities. His research focuses primarily on how public service delivery can be improved through reform initiatives and improvements in political conditions. In particular, he is interested in how technological progress and a robust media can underwrite such improvements.
Organized by the Economics Department