Stanford scientists develop tool that predicts virality of photos on Facebook
A recent report claims that a group of Stanford computer researchers have developed an algorithm which can predict if a photo will go ‘viral’ on Facebook.
Stanford researchers said that the rate of multiple sharing or cascades is the key to predict which of the many millions of photos on Facebook will go viral. The team will present their study at the upcoming International World Wide Web Conference.
According to the data provided by Facebook to university scientists, only 1 out of every 20 pictures posted on the social network manages to get a single share, while just 1 in 4,000 photos get more than 500 shares.
The Stanford team analysed 150,000 Facebook photos, which were shared at least 5 times, in order to formulate the algorithm. It was observed that, at any given point in a cascade, there was a 50-50 possibility of the photographs earning double the number of shares. The scientists looked out for variables like the rate and speed at which photos were shared, and the structure of sharing.
After analysing several criteria the researchers were able to accurately predict doubling events almost 80 percent of the time. The researchers claimed that speed of sharing was the best predictor as by analysing how quickly a cascade unfolded.
The scientists also noted how a photo was shared or the structure of the cascade, to be the next best predictive factor with an accuracy of 67.1 percent at predicting doubling.
“It wasn’t clear whether information cascades could be predicted because they happen so rarely,” said Jure Leskovec, assistant professor of computer science. “Slow, persistent cascades don’t really double in size.”
“Even if you have the best cat picture ever, it could work for your network, but not for my boring academic friends,” Leskovec says. “You have to understand your network.”