Flickr Can Predict Where You’ll Be in the Future

The photos you uploaded to Flickr of your summer vacation may allow strangers to predict where you are at a given time in the future, a new study shows.

Researchers at University College in England analyzed timestamp and location metadata of images taken in the UK and uploaded to the photo sharing site between 2007 and 2013. The study, published Wednesday in Royal Society Open Science, moved 8 million images uploaded by 16,000 into a separate database for analysis. Using a computer algorithm, researchers ran the geotag information from each individual camera through an algorithm to predict where users would be in the future.

They found the data showed individuals showed a Lévy-flight pattern, meaning they moved around locally, in a relatively small area, and later to larger movements to a distant area.

"We propose that the movement of Flickr users, observed by collecting geo-tagged metadata, is driven by unobserved factors," the researchers wrote. "For example, one individual might reside in Bristol, visit relatives in Suffolk, and spend holidays in northern Wales."

The results were consistent with official data on travel patterns 92 percent of the time, suggesting Flickr information could be used to accurately predict this movement.

The researchers suggested mining data from online sources like these could someday model and predict human mobility on a large scale, with potential uses in urban planning and transportation development.