Lunch meetings at Takara Sushi are a regular occurrence for the PoliticIt group, a politically-themed team comprised of various USU students and professors. The weekly meeting gives them a chance to go over business while enjoying good friends and food. At one particular meeting last August, they brainstormed and chatted while dining on various sushi rolls. Amidst the friendly chatter came an idea for a website that tracked and profiled politicians. With the 2012 presidential race coming up, the timing couldn’t be better, they thought.
From there, the idea snowballed. The group gave each other assignments dealing with site content, conceptual design and research. Excitement filled the restaurant’s air as ideas rolled off the tongues of the creators of a company that specializes in making prognostications in elections and other political events.
After lunch ended, they got to work. By Sept. 2011, only a month after the initial meeting, PoliticIt.com was born. Creators include USU Business Management department head John Johnson, USU alumni Josh Light and Britney Johnson, USU graduate student Sterling Morris and USU undergrads Lauren Johnson and Shai McDonald. The site tracks social media buzz surrounding political candidates. By measuring social media, PoliticIt has found a way to predict the outcomes of political races in three ways: be predicting polls before they come out; by tracking data on how people feel about politicians over time; and also by being a “one-stop shop,” Light said, for people to stay up to date about the presidential candidates in the current race.
“Digital influence is positively correlated to poll performance,” Light said in an interview with Utah Public Radio. Take Twitter. Light said the PoliticIt team looks at all the tweets and other online comments surrounding candidates and decipher between the positive and negative comments. This contributes to a candidate’s “It Score”, which is given to a candidate based on what people do and say on the internet. If what is said is found positive, the It Score will increase. If it is negative, it will decrease.
It Scores are calculated in real time and changes reflect new “developments across the internet concerning a politician,” as described on the PoliticIt website. The team uses “sophisticated” software called neural networks, Light said. This software uses genetic algorithms that compile all of the data found in social media to predict the outcomes of politicians. The PoliticIt team strives to provide objective data. Objectivity allows people to trust the data, knowing that there are no personal agendas polluting information that is given. “We are all about eliminating asymmetric information in political markets,” Light said. “We want people to vote on issues and things that are relevant, not on someone’s hair or name.” He said that the goal is to make politics and the internet more “transparent.”
The info for the site is not just pulled from Facebook and Twitter; it is gathered from Intrade, Wikipedia hits, YouTube views, Klout, Yahoo news stories and Google trends. This enables the site to feature a more comprehensive profile on candidates and contain a larger database to pull their predictions from, Light explained. That includes a “new age,” called stat 2.0, where large amounts of data can be collected at a cheap price.
“We’re entering into a new age of statistics,” Light said. “The only cost is storing the data. We’re only in politics and data, but there are so many applications.” PoliticIt has been recognized many times for their work, including a top five placement in the Intel Innovators list and a win in the USU Jon M. Huntsman School of Business Opportunity Quest. PoliticIt is receiving funding from Angel Investors to continue operations. However, in the near future they hope to be able to sell their software to politicians and to expand the website after finishing coverage on all 50 states.