X
Innovation

How Twitter knows when you nap

By looking at lulls in Twitter activity, the site's engineers can tell when people sleep and when they wake.
Written by Sarah Korones, Contributor

Think you’re not producing any data when you’re fast asleep? Think again.

Two engineers at Twitter have discovered a way to determine when users are sleeping and the insights from this intelligence could prove invaluable to advertisers hoping to maximize exposure.

After analyzing the social network’s data from four major cities around the world, Twitter engineers Miguel Rios and Jimmy Lin were able to determine when people wake, when they sleep and whether or not they tweet at work. The findings are based on lulls in each city’s Twitter activity—when people weren’t busy firing off tweets, the engineers figured they were asleep.

The visualization of the data above reveals some key differences between those in New York City, Istanbul, Tokyo and Sao Paulo. While people in New York, Istanbul and Sao Paulo all stay up later in the summer, those in the Big Apple tend to burn the midnight oil all year round. Citizens of Sao Paulo sleep later and seem to take a midday siesta—their tweets drop off noticeably after lunch.

Unlike those in the other three cities, New Yorkers are just as likely to tweet during the day as they are in the evening. Residents of Tokyo, on the other hand, mostly confine their Twitter activity to before or after work.

The intelligence could help outside companies market their products on Twitter—informing them of the best times to place advertisements and making sure they’re not airing ads to an empty room.

“It [the intelligence] can inform when to serve up advertisements on the site and potentially for what kinds of products,” the New York Times reports. “The data at the disposal of social media companies are, in principle, valuable to other organizations, from soda pop companies to politicians to police agencies that seek to understand human behavior.”

Twitters Knows When You Sleep, And More [NYT]

Image: Twitter Engineering Blog

This post was originally published on Smartplanet.com

Editorial standards