Illuminating 2016 is a computational journalism project that will empower journalists covering the 2016 presidential campaign. Political reporters must cover not only stump speeches, campaign events, and TV ads, but also what is happening on social media. Covering it increases transparency and accountability of the campaigns, and is a way to take the pulse of the electorate. The sheer volume of information that flows through social media, however, makes it challenging to report accurately and comprehensively. Our goal is to help journalists in that important work by providing a useable yet comprehensive summary of the content and sentiment—that goes beyond counting likes or retweets. Illuminating 2016 will enable political journalists an insightful yet accessible summation of the important political conversation online.
This project is supported by the Knight Foundation, the Tow Center for Digital Journalism at Columbia University, the Center for Computational and Data Sciences and the BITS Lab at the School of Information Studies at Syracuse University
The Illuminating 2016 project has several goals. First, as a research project we seek to understand what indicators on social media can be used to determine support for presidential candidates. Journalists tend to look at the most easily observable metrics, such as followership rates. Yet, there are likely better indicators that determine electoral success. Second, as an applied project, we seek to understand what political journalists need to best report on elections, factoring in the complex social media landscape. That is, we have the opportunity to work with journalists through the primaries, conventions, and general election to provide analysis and data visualizations most useful to their reporting.
To achieve these objectives, we are collecting the Facebook and Twitter messages and images of major party presidential candidates. We are also collecting Facebook comments, and retweets and mentions on Twitter. We are building an automated system to tag the types, topics, and sentiment of candidate messages, and will do the same for the public’s conversation. Critical for this work is to interview and observe about 30 political journalists to understand their needs for reporting. Interviews will inform the design of a website we will build in the spring to provide data visualizations and automated reports.
In our early conversations with journalists, they highlight challenges in reporting on the social media of campaigns because of the volume, multiplicity of platforms, and technical challenges in collecting, aggregating, and analyzing messages by the candidates and the public. Journalists express genuine interest in knowing what are indicators of probable success. For example, should the massive following that Donald Trump has on Twitter suggest he will be the eventual nominee? Hence, one significant and timely benefit of this project is to support journalists in their reporting of the 2016 election cycle, and in future election cycles. We also expect to present at conferences and to publish the use and design of the Illuminating 2016 website using design-science and participatory design principles to assist practitioners.
Second, we will advance research. We expect to publish two to three significant conference papers and peer-reviewed journal articles. These will examine the change in campaign messaging over time on social media, the nature or extent to which presidential candidates drive the policy discussion that the public has during the campaign, the types of campaign messages the public interacts with and spreads the most, the ways that major events, such as debates or political gaffes, are talked about or go viral and how that shapes the discussion, as well as analysis about political fragmentation and the extent to which the public in 2016 talked only with like-minded people. These papers will advance basic knowledge and also aid journalists in considering the role of social media in campaigns.
Third, we are training masters and doctoral students in the techniques that support computational journalism, which is an increasingly needed skill in newsrooms.
To cite the Illuminating 2016 website, e.g. data, images:
Stromer-Galley, J., Hemsley, J., Tanupabrungsun, S., Zhang, F., Rossini, P., Bryant, L., McCracken, N., Hegde, Y., Semaan, B., Jackson, S., Boichak, O., Li, Y., Harandi, M., Robinson, J. (2016). Illuminating 2016 Project. http://illuminating.ischool.syr.edu
To cite the categories and the machine learning work that created the categories:
Zhang, F., Stromer-Galley, J., Tanupabrungsun, S., Hegde, Y., McCracken, N., Hemsley, J. (2017). Understanding Discourse Acts: Political Campaign Messages Classification on Facebook and Twitter. Proceedings of the 2017 International Conference on Social Computing, Behavioral-Cultural Modeling & Prediction and Behavior Representation in Modeling and Simulation (SBP'17), George Washington University, DC, July 5-8, 2017.
To cite the overall methods and collection of social media messages for the Illuminating 2016 project:
Hemsley, J., Stromer-Galley, J., Tanupabrungsun, S., Hegde, Y., Zhang, F., McCracken, N. (2017). Collection and Classification of Illuminating 2016 Social Media Data. http://illuminating.ischool.syr.edu/blog