Campus Connection

UWL professors, student build mathematical model to compare Twitter “outbreaks”

[caption id="attachment_41736" align="alignright" width="300"] Math professors, students compare discussion on football scandal to spread of infectious disease[/caption] More than five months ago, a report that the New England Patriots cheated in one of their NFL playoff games surfaced. It was just the event UWL Professors Eric Eager and James Peirce, and student Megan Eberle were waiting for. The trio had built a model to compare the spread of information to that of an infectious disease. “Information within itself has the epidemiology structure,” notes Eager. They analyzed a set of data collected from Twitter. Every user was considered susceptible to catch the disease, and every message, or tweet, sent out about Deflategate meant someone was infected. “Think about it like a cold hitting campus,” explains Peirce. “Everyone gets sick. Then after a certain time people get better. Eventually people stop talking about the popular stories in the same way.” The data showed an average group of 1,000 users were able to get thousands more to tweet about Deflategate. And the average individual would be “infected” – or continue talking about it – for about three days.

Comparing “outbreaks”: some are shorter

The researchers also applied a slightly modified version of their model to the announcement of Hillary Clinton’s presidential campaign and the stories surrounding the Baltimore riots resulting from the death of Freddie Gray. Clinton’s announcement had more total tweets at its peak than Deflategate, however most people were done talking about it less than 16 hours later. “We can write that off because there are multiple people announcing candidacy for president and because of all the buildup to the announcement,” says Eager. The results from the Baltimore Riots “revealed something far more disturbing,” notes Eager. “While Deflategate followed the disease model right, Freddie Gray was like an earthquake,” says Perice. “There were far more initial tweets, but it barely had any staying power.” The researchers believe people are tired of stories about questionable police action. “However, one would hope for the same such fatigue over relatively meaningless NFL stories, which we have shown is not the case,” explains the researchers in the paper. “We care about the NFL more because it gives us an escape from these things,” says Eager. “The story from Baltimore is far more real.”

Classroom application intrigues students

The idea to look at Deflategate came from Eberle, a student heading into her sophomore year at UWL. The professors say other students are jealous she got to study her own problem from start to finish. “The problem wasn’t made up. It was authentic,” says Peirce. Eberle participated in Eagle Apprentice, a program that funds 50 students to collaborate with faculty members on research. “She now has a unique toolset to solve some cool problems, and will have a far more interesting perspective going through her courses,” says Eager.” Eager and Perice hope to take on another Eagle Apprentice this fall. They also hope classes will use their research as an example of real-world application for mathematics. “The idea of gathering data from a new story catches students imaginations,” notes Peirce. “It gives them a chance to apply some of the advanced techniques they’ve learned and be excited about it.”