Can Studying Tweets Lead to Faster Earthquake Recovery?

What's This?

A boy searches through debris after multiple earthquakes toppled this building in Nicaragua.

Image: Esteban Felix/Associated Press

By Kurt Wagner2014-05-02 18:58:57 UTC

Twitter has always been a place to talk about what's happening around you — even if what's happening is an earthquake.

In March, an early morning earthquake in Los Angeles resulted in a Twitter storm of sleepy, yet humorous, tweets. That type of real-time sharing is the basis of new research out of Stanford University, where researchers recently tested Twitter's ability to accurately detect ground-shaking intensity, a metric displayed using visual graphs known as ShakeMaps.

See also: 10 Must-Follow Breaking News Accounts on Twitter

In other words, the real-time tweets led scientists to wonder — can those tweeting about an Earthquake help researchers learn more about it?

According to the Stanford researchers, they can. Reza Zadeh, a lecturer at the Institute of Computational and Mathematical Engineering at Stanford University, along with two PhD candidates, found that including data pulled from user tweets alongside traditional earthquake measurement data resulted in more accurate ShakeMaps than those without the Twitter data.

The trio studied tweets from Twitter users in Japan following earthquakes of a magnitude 6 or higher in 2011 and 2012. The tweets, which were only included if they were geo-taggged with the user's location, were examined if they mentioned earthquake-related terms. Then the researchers looked at other elements of the tweets, like the number of times these terms were mentioned, how many exclamation points a user included, or if they added emoji.

The tweets helped give researchers an idea of how the earthquake impacted certain areas. This type of information is usually added to ShakeMaps and geological readings gathered by scientists, but it normally takes time to collect. According to the study:

As time progresses, the ShakeMap is continually updated as new information becomes available, including “did you feel it?” data, which are qualitative first-hand accounts of the earthquake collected via online surveys. While this qualitative information is extremely informative, it can take anywhere from hours to days to weeks to collect and is therefore not typically included in very early shaking intensity estimates.

Following a separate 5.1-magnitude earthquake in Los Angeles in March, tweets mentioning the word "earthquake" jumped from roughly 50 tweets-per-minute, to more than 24,000 TPM immediately following the quake, according to data provided to Mashable by Twitter. The quake hit at 9:09 p.m. PT, when there were 50 TPM mentioning "earthquake." Between 9:10 and 9:11 p.m., "earthquake" was tweeted 14,500 times. Between 9:11 and 9:12 p.m., that number rose to 24,500 TPM.

The point? Twitter users can provide seismologists with thousands of first-hand accounts within minutes, meaning, potentially, more accurate ShakeMaps in a shorter period of time.

The benefits of more accurate maps mainly effect earthquake recovery. If rescue officials can determine which areas were hit the hardest based, partially, on tweets, they can more accurately triage aid and determine where and when emergency vehicles should be deployed.

Tweets are not currently part of the formula for generating ShakeMaps, but they are being considered. The U.S Geological Survey released a report in 2011 citing earthquake-related Twitter activity as a potential supplement to the organization's existing measurement tools. Similar to the Stanford study, the USGS found that the immediacy and public availability of tweets could be a benefit to researchers and rescue workers.

There are also disadvantages of using Twitter, the study found. For example, when first-hand accounts are taken via email survey, experts can include "customized magnitude ranges and regions" to guide user responses. Tweets may not be that detailed. As the USGS put it: "The main downside is the difficulty in providing detail and actionable information in 140 characters. This restriction results in numerous tweets of marginal information value."

Twitter is already used widely for dispersing safety information in times of crisis — including earthquakes. The USGS tweets out information from multiple Twitter accounts, and other agencies worldwide like Natural Resources Canada and the Indonesian Agency for Meteorology, Climatology and Geophysics also use Twitter to alert citizens of potential earthquake threats.

The addition of Twitter data to hone ShakeMaps would only makes sense in areas where Twitter use is heavy and the event of earthquake would generate a substantial increase in tweets, which does seem to play into Twitter's strengths, to an extent. The social network has large user bases in both the United States (Twitter HQ is right on the San Andreas fault line, along with the rest of Silicon Valley), and Japan, another hotbed for seismic activity.

Have something to add to this story? Share it in the comments.

Topics: Climate, Earthquake, fault lines, shakemaps, Social Media, Twitter, USGS