Stanford researchers discover predictors for viral photos

A team of computer scientists, including a Stanford professor and doctoral student, recently identified some indicators that can assist in predicting which photos on Facebook will go viral.

After analyzing 150,000 Facebook photos that had been shared at least five times, the group was able to predict with 80 percent accuracy whether or not a photo would double in shares.

The group found that the speed of sharing was the best predictor of whether the photos would double in the number of shares. Additionally, the way that a photo was shared was also a powerful predictor. Photos that were shared among different friendship networks and fan groups were more likely to go viral than photos that were only shared within a specific network.

The researchers, however, have not been unable to determine what content in the photo actually makes a photo go viral.

“Even if you have the best cat picture ever, it could work for your network, but not for my boring academic friends,” said Jure Leskovec, assistant professor of computer science, to the Stanford News Service. “You have to understand your network.”

 

Andrew Vogeley

About Andrew Vogeley

Andrew Vogeley ‘17 is a sophomore majoring in political science. Andrew hails from the great state of Texas (and he’ll be sure to let you know it) and serves as a news desk editor, covering the different student groups on campus. Besides editing and writing for The Daily, Andrew is President of RUF, a Christian fellowship group. To contact Andrew, email him at avogeley ‘at’ stanford.edu
Advertisment ad adsense adlogger