Using only tweets, scientists can now detect earthquakes in 29 seconds flat

You need to get on Twitter.
You need to get on Twitter.
Image: Reuters/Ahmad Masood
By

Did you first hear about the devastating earthquake that struck Sichuan on Twitter? You’d be surprised at how many did, as the social-media platform was actually faster at reporting the earthquake than the US government organization tasked with monitoring such events.

The US Geological Survey (USGS) has 2,000 earthquake sensors but the vast majority of these are based in the US. This limits the USGS’ ability to monitor earthquakes in the rest of the world. To cover its blind spots, USGS has teamed up with Twitter.

Millions of people use Twitter to report earthquakes, but the data needs to be fine-tuned for it to be useful. USGS analyzed these tweets and found that those tweeting about earthquakes kept their tweets short. They also realized that those tweeting links were less likely to be users experiencing the earthquake firsthand.

So, the USGS decided to filter out tweets with more than seven words and those that contained links. These filtered tweets proved to be effective at monitoring earthquakes globally. Now, when a number of people start tweeting about an earthquake in the area, the USGS gets an alert.

It only took one minute and 20 seconds—from just 14 tweets—to be alerted of an earthquake aftershock in Chile.

In 2014, the USGS was alerted to the earthquake in Napa, California in 29 seconds using Twitter data, the company said on its blog post. This data also allows the USGS to improve their own detection system and acts as a secondary check, so if a sensor detects an earthquake in a densely populated area but no-one is tweeting about it, then the USGS knows it’s a false alarm.