November 21, 2017

Syracuse iSchool Social Media Tool Used to Track Elections

BITS Lab at Syracuse University iSchool Github graphicResearchers at the Syracuse University School of Information Studies (iSchool) are using the open-source Social Media Tracker, Analyzer, and Collector Toolkit at Syracuse (STACKS) to collect and analyze social media posts and traffic related to the 2016 presidential candidates as part of an interdisciplinary digital politics project, according to a campus publication.

“Campaigns are now using social media as one of the many tools for persuading people to vote and for attacking opponents,” Jenny Stromer-Galley, iSchool professor and author of Presidential Campaigning in the Internet Age, told Syracuse University Magazine. “Social media, more than TV advertising, also enables a campaign to encourage supporters to work for it.”

Created by lead developer Jeffery Hemsley, iSchool Assistant Professor and co-author of Going Viral, winner of the Association of Information Science and Technology’s 2014 Best Book Award, STACKS currently works with Twitter’s streaming API. The capability to collect data from Facebook and the Twitter search API is under development by Hemsley and others at the iSchool’s Behavior, Information, Technology, and Society Laboratory (BITS Lab).

“Anything that matches our search terms, Twitter streams back to us, and we maintain a connection that’s always on,” Hemsley told LJ. “Currently we’re collecting data about the upcoming presidential election. So, if we’re collecting tweets sent by Donald Trump, anybody that uses [his] handle in a tweet, we also get too. We also get any tweets that any presidential candidate retweets, or if they @mention somebody, or someone @mentions them.”

Hemsley, whose pre-academic career included software engineering positions at companies such as Symantec and Autodesk, began developing STACKS to collect data for research projects while working on his MLIS and PhD in Information Science at the University of Washington’s iSchool. During that time, Hemsley said, he became interested in the concept of social media as a democratizing force.

“We’re no longer limited in terms of the information that we get through just a few gatekeepers,” he explained. “The traditional media makes decisions about what we hear about. But now, you can share information, so there’s the potential for us to hear about things that we wouldn’t otherwise hear about. That’s why virality is so important to me. Every time we share a message, every time you see a video and think, ‘Hey, I’m going to share that with my friends,’ there’s some way in which you are voting for content. You’re saying, ‘This is important enough for us to talk about. Important enough to share.'”

In addition, Hemsley noted that in many ways, social media has displaced diaries, journals, and letters as ways that people write about their lives and thoughts.

“There was a time when people kept diaries, and…we can look back at them as historical documents. Now, we have social media data, and people are doing kind of a similar thing. They’re documenting their days, they’re documenting what’s important to them. Virality is, at its heart, in my mind, a way in which the crowd is voting, and saying this is what we should be paying attention to.”

While using a tool like STACKS to collect and study the musings, retorts, and marketing messages sent and received by a field of 2016 presidential candidates certainly sounds like serious academic work, Hemsley also defends the relevance of social media’s apparently frivolous side—Justin Beiber’s 30,000 tweets to almost 70 million followers, or the popularity of anything cat related, for example.

“If a million people are talking about Justin Bieber, he’s reflecting something that is interesting or relevant about the culture today,” Hemsley said. “Some people might feel like tweeting about cats or Justin Bieber is trivial, but if enough people are doing it, there’s something there. It’s reflecting something that society thinks is important. Maybe it’s our materialistic culture, maybe it’s ‘cuteness’ which is a basic part of humanity, or maybe it’s indicative of the fact that our culture, currently, is very celebrity driven. We look at celebrities. Why do we look at celebrities? Some people have said it’s because we think our own lives aren’t very interesting. Or maybe it’s because they’re doing something that we wish we could do. I don’t know. That’s not my area of specialty…. But Twitter is a reflection of what some portion of society thinks is important.”

Hemsley did note that social media research tends to collect information about a subset of the population that is very media oriented, and that different social media platforms such as Twitter, Instagram, and Facebook each attract a unique mix of age groups and demographics. Still, he suggested that studying online opinions and reactions, perhaps someday mining the Twitter Archive at the Library of Congress (LC)—which has been stalled as LC deals with technical and policy-related challenges regarding how it might provide access to every tweet posted since the service’s inception—could have significant implications in the future.

“A government, for example, that is mining social media could use that for good purposes, like figuring out what policies or projects and things are working and what isn’t,” he said. “Where could government help us, for instance?”

In contrast to the data mining potential of LC’s Twitter archive project, STACKS is more appropriate for targeted research projects focused on specific people or topics, such as the current field of presidential candidates. Written primarily in Python, STACKS uses a MongoDB NoSQL database, which handles unstructured and semi-structured data more easily than a traditional structured query language (SQL) relational database. The open source program is available on GitHub, and will run on Mac OSX or any Linux environment. Interested users would also need to install MongoDB, which is open source, and Python.

Matt Enis About Matt Enis

Matt Enis (menis@mediasourceinc.com; @matthewenis on Twitter) is Senior Editor, Technology for Library Journal.

Share