Big Data 2012: Atigeo Analyzes Campaign Posts and Tweets

Aug 22, 2012

Article ImageEvery four years brings a new opportunity to analyze the U.S. presidential election using the most state-of-the-art technological tools. From the first known public opinion poll, which was conducted during the 1824 presidential election between Andrew Jackson and John Quincy Adams, to Nate Silver's sophisticated statistical prediction algorithms in 2008 on The New York Times' FiveThirtyEight blog, analysts, pundits, and informed voters alike have long sought out the savviest ways to make use of the data at their disposal during presidential campaigns.

The 2012 cycle marks the inaugural election during which "big data" analysis tools are widely available to examine where the Republican and Democratic presidential candidates stand on the major issues of the campaign. The technology company Atigeo is putting its big data cloud software platform, xPatterns, to work in a free online analytical tool that it hopes voters will use to sift through the cacophony of campaign commentary and make sense of the candidates' stances on the issues.

The tool, which is actually a pair of websites called Red2012Blue and Blue2012Red, generates lists of links to posts on the Facebook pages, Twitter accounts, and RSS feeds by the campaigns of President Barack Obama and former Massachusetts Governor Mitt Romney, as well as those of the Democratic National Committee and the Republican National Committee.

By searching for, or clicking on, a campaign issue such as "healthcare," users of the site are presented with side-by-side lists of the most recent social media postings by the parties and candidates. (The lists are presented under the red, white and blue donkey and elephant icons, which are the respective symbols of the Democrats and the Republicans.) Users can then click the links to go directly to the page where the content is located.

Atigeo COO and chief security officer Chris Burgess says the tool came about in part because of his own experience in a politically active family. His parents were advocates for overturning overly restrictive voting laws, and through their influence he came to view voting as "the most important right that we have as American citizens," he said.

"I got tired, personally, of pundits and analysts taking the candidates' words and spinning them their own ways on television and in the media," Burgess said. "So we talked amongst ourselves and asked where to get that information directly from the candidates. These days, it's Facebook, Twitter, and the RSS feeds of the candidates and their parties. Why not take that information, which is in its rawest state, and use our big data analytics tool and see what happens based on that content."

So, why two separate websites? Both offer the same information, but Burgess found, through unscientific but nevertheless persuasive conversations with colleagues and family, that users of the tool inferred a political bias based on which political color came first (red representing Republicans, and blue representing Democrats). To avoid any accusation of a political tilt, Burgess and Atigeo decided to make the tool available at two different URLs with inverse names and layouts.

"So far, the two sites are running neck-and-neck in terms of traffic," Burgess says. "And about thirty percent of visitors are visiting both sites."

The Red2012Blue/Blue2012Red project also afforded Atigeo the opportunity to demonstrate the capabilities of its big data analytics platform. The xPatterns tool assigns relevance scores and generates domain concepts that allow users of the site to search topics in clusters. Also available on the sites is access to Atigeo's visualization tool, which allows for easy navigation (with 3D zoom-and-drag capabilities) of non-obvious relevant concepts and documents relating to the original search query.

Burgess is convinced that social media is on its way to becoming the predominant tool for candidates, campaigns, and parties to communicate their messages, and he sees that channel of communication - that is, direct to the voter, as opposed to filtered through the media - as an important source of information for a well-informed electorate. And with all of that information out there, it'll take smart, sophisticated big data tools to sort through it.

("Obama Romney Street Sign" courtesy of Shutterstock.)