Using machine learning to uncover the roots of Trump online support

Why do Trump supporters support Trump? How was his political campaign able to attract a large support, ultimately reshaping the U.S. political landscape? In a new study, a team of ISI Foundation researchers — Corrado Monti, Gianmarco De Francisci Morales, Francesco Bonchi — together with Joan Massachs, a student from the Polytechnic University of Catalonia, used machine learning and social media to give some data-driven answers.

By extracting data from the social media Reddit, they investigate The Donald, one of the largest online communities of Trump supporters. “We used machine learning algorithms to gauge which users publicly support Trump in 2016, but only by using data from their actions in 2012, four years earlier. In this way, we train an algorithm to recognize what early signals are more important, telling us something about which characteristics were more prominent in early Trump supporters. For instance, we found that participation in male-dedicated communities, conservative and libertarian groups, are among the strongest predictors, along with certain sets of interest (like firearms and entrepreneurship). We also found links with conspiracy theories groups, as well as a tendency to receive negative feedback from the rest of the Reddit community".

This study is going to be presented at the 12th ACM Web Science Conference (July 6th – July 10th), and it is available as a pre print. Online communities have been used as propaganda machines by many political campaigns, and it is therefore important to understand how they are created and how are they able to grow. The investigation of how social media affects public opinion is of immediate importance.

“Roots of Trumpism: Homophily and Social Feedback in Donald Trump Support on Reddit”, Joan Massachs, Corrado Monti, Gianmarco De Francisci Morales, Francesco Bonchi. Link