Stanford University researchers have developed a computer algorithm that can predict which photos will likely go viral on Facebook.
In a report being presented at the International World Wide Web Conference in Seoul, South Korea, Jure Leskovec, Stanford doctoral student Justin Cheng, Facebook researchers Lada Adamic and P Alex Dow, and Cornell University computer scientist Jon Kleinberg, describe how they created an algorithm capable of accurately predicting (8 out of 10 times) which Facebook posted photos would go viral.
Statistically, based on data provided by Facebook scientists, only one in 20 photos posted on the social network gets shared maybe once. And just one in 4,000 gets more than 500 shares.
While reviewing what they called photo cascades, researchers studied how quickly a photo was seen and shared – garnering clues as to how a photo goes from being relatively obscure to prolific on cyberspace.
The term ‘cascades’ is used to describe photos or videos being shared multiple times.
The team began by analyzing 150,000 Facebook photos, each of which had been shared at least five times. The data was stripped of identifiers to protect user privacy.
A preliminary analysis of test photos revealed that, at any given point in a cascade, there was a 50-50 chance that the number of shares would double. Variables were assessed in order to determine when doubling events would likely occur, explains The Stanford Daily.
After factoring other criteria, the scientists were able to accurately predict doubling events almost 80 percent of the time. The speed of sharing was the best predictor. Their algorithm became more accurate the more times a photo was shared; 88 percent for photos shared hundreds of times.
But what aspects of a photo drives it to go viral? Alas, other than being shared on multiple platforms, the researchers found no simple trick to ensure widespread sharing.