Understanding election coverage using Big data

Researchers from our Department of Media and Communication, in collaboration with the University of Bristol, have for the first time analysed mass media coverage to understand how elections are covered by the press.

An analysis of more than 130,000 online news articles has revealed how the 2012 US Presidential election played out in the media, with the results recently published in the journal Big Data & Society.

In the first study where political positions have been automatically obtained from a very large amount of online news using a semantic graph and linking text to identified noun phrases and verbs, the researchers confirmed that the range of political positions can be reliably uncovered from the set of claims attributed to each actor.

Dr Giuseppe Veltri from the Department of Media and Communication said: “We live in an information dense and highly mediated society where it is difficult to make sense of major events of public life. The application of Big data methods, as in this study, can help us make sense of the complexity of public events such as elections and their intricate unfolding of alliances, endorsements and inconsistencies among actors and issues.”

This method of identifying subject-verb-object triplets was used as a building block for the network of analysis which has never been applied to a real-world dataset on this scale. This exemplifies how big data and computational research can help in understanding the dynamics of election campaigns.

• The paper, ‘Automated analysis of the US Presidential elections using Big Data and network analysis’ published in Big Data & Society, 2015; 2 (1) DOI: 10.1177/2053951715572916