Newswise — As Iowa surges toward the presidential caucuses, University of Iowa computer science professors M. Zubair Shafiq and Padmini Srinivasan and political science professor G.R. Boynton have launched a political news aggregator, leftrightpolitics.com, to track public sentiment during the election season.
The website’s features include graphs tracking public sentiment of the leading candidates from both the Republican and Democratic parties, a feed of the most talked-about news, and a more detailed breakdown of each candidate, including the publics’ perception of various attributes (whether the candidate is friendly, forceful, pacifist, intellectually brilliant, etc.), and a breakdown of who it is that’s doing the talking about candidates—Republicans, Democrats, or independents.
In front of every news article on the site is a bias meter showing whether the article leans toward a Republican bias, a Democratic bias, or remains neutral. This algorithm is based on the estimated political leanings of people retweeting the article. Other websites claim to sort articles based on bias, but they rely on humans to manually report that bias, Shafiq says, so a truly accurate bias rating requires an extremely large number of people reporting.
For example, AllSides is a competitor that also reports the bias of particular news sources. It shows The New York Times as leaning Democratic, Fox News as very Republican, and CNN as centered. Shafiq says his algorithms provide a more nuanced picture. The New York Times and CNN both have a spread of biases, and though Fox News is largely Republican, there are a number of more moderate articles.
Shafiq hopes his website will help educate people so that when they go out to vote they have enough information to make an informed choice.
“We want to bring diverse news so you don’t get stuck in your own echo chamber,” Shafiq says.
A unique feature of the technology is that it can be applied to articles in any language because it is based on article shares, rather than the content itself.
So far, the website has unearthed a few interesting trends, such as the fact that on social media, people tend to talk about nontraditional news sources, rather than more established and mainstream organizations like The Washington Post or The New York Times.
Additionally, the candidate breakdown shows that often, the greatest volume of chatter about a candidate is from his or her opposition. Bernie Sanders, for example, is discussed much more by Republicans criticizing his views than by Democrats in favor of his ideas. In traditional politics, voters can only attend primaries for the party with which they are affiliated, but social media breaks down those barriers and allows conversation across party lines.
The website has a Facebook page, Twitter account, and a blog at http://www.medium.com/@LeftRightPolitics, where G.R. Boynton, UI professor of political science, will comment on interesting trends as the election season progresses.In the future, Shafiq hopes to apply the technology to other potential fields of commercialization, such as product reviews. He says that companies may be interested in knowing whether people are saying positive or negatives things about their products or services.
Shafiq, Boynton, and Srinivasan received gap funding through the UI Ventures program in the Office of the Vice President of Research and Economic Development for this project last semester. The funding helped them turn their algorithms into a commercial venture and enabled them to hire web developers to create the website featuring a real-time service that can be viewed by anyone.
Shafiq says the value of gap funding goes beyond the money; the motivation and ideas provided by the team at the UI Research Foundation and UI Ventures are also crucial.
“As an academic, I’d never launched a company,” Shafiq says. “They helped us go from an academic mindset to a business mindset and helped us think of unique ways to commercialize these technologies.”
RSS