skip to main content

Listen

Read

Watch

Schedules

Programs

Events

Give

Account

Donation Heart Ribbon

Social Media Analysis And The San Diego Mayor’s Race

Evening Edition

Aired 11/8/13 on KPBS News.

To better predict San Diego’s upcoming mayoral election, a professor at San Diego State University has tapped into the seediest and most vile underbelly of American politics.

To better predict San Diego’s upcoming mayoral election, a professor at San Diego State University has tapped into the seediest and most vile underbelly of American politics.

Twitter.

Last month, Professor Ming-Hsiang Tsou of SDSU’s geography department, along with colleagues and research students, created “ElectionPath.” It’s an interactive web app that analyzes the Twitter buzz surrounding Mike Aguirre, David Alvarez, Kevin Faulconer and Nathan Fletcher -- the top four men vying to become San Diego’s next mayor.

Tsou and his colleagues believe social media analysis -- which includes harvesting each candidate’s daily change in followers, mentions, retweets and other categories -- will eventually serve as an inexpensive and more accurate predictor of electoral outcomes than old-school polling methods.

“I believe this is the future for all the political campaigns,” Tsou said.

Tsou is not alone. After analyzing more than half a billion tweets related to candidates in the 2010 Congressional elections, a team of academics in Indiana found Twitter traffic predicted the winner in 404 out of the 435 races. Yet several critics have pointed out flaws in the team's quantitative methods, including a notable counterpoint from the Rothenberg Political Report:

“Since House re-election rates have been over 90 percent in 19 of the past 23 elections, you don’t need polls or tweet counts to predict the overwhelming majority of race outcomes. In most cases, all you need to know is incumbency (or the district’s political bent) and the candidates’ parties to predict who will win.”

Tsou was not completely familiar with the Indiana study or its criticisms, but did note the inherent flaws of “big data.”

“It's very important to identify the white noise -- the error -- and create a filter,” he said.

Tsou likened the social media sampling to weather predictions.

You don’t need measurements from every location in a city to predict tomorrow’s weather, you need good sampling skills and a handful of strategically-placed thermometers, he said.

In addition to the Twitter analysis, Tsou and colleagues are working on other ways to incorporate social media analysis with real world trends, such as predicting the scope of the next flu season. He told inewsource his team has been in communication with the Centers for Disease Control and Prevention, which he sees as a positive sign for the future of his endeavor.

To use Professor Tsou's application, click here.

Disclaimer: inewsource is a partner of SDSU and a supporter of emerging technologies created by our colleagues. We are hosting Tsou’s new application on our website with a strong disclaimer: we are not social scientists. We believe this technology is too new to say definitively whether it is a bellwether for election results. But we’re interested in following Tsou’s work, and believe our readers may be, as well.

Questions, tips or suggestions? Email reporter Brad Racino at bradracino@inewsource.org or call him at 619.594.3569. Follow Brad on Twitter at @BradRacino.

We've upgraded to a better commenting experience!
Log in with your social profile or create a Disqus account.

Please stay on topic and be as concise as possible. Leaving a comment means you agree to our Community Discussion Rules. We like civilized discourse. We don't like spam, lying, profanity, harassment or personal attacks.

comments powered by Disqus