Modeling “wordquakes”

Several researchers suggest that certain words on the Internet are used in patterns similar to those of earthquakes:

News tends to move quickly through the public consciousness, noted physicist Peter Klimek of the Medical University of Vienna and colleagues in a paper posted on Readers usually absorb a story, discuss it with their friends, and then forget it. But some events send lasting reverberations through society, changing opinions and even governments.

“It is tempting to see such media events as a human, social excitable medium,” wrote Klimek’s team. “One may view them as a social analog to earthquakes.”…

Events that came from outside the blogosphere also seemed to exhibit aftershocks that line up with Omori’s law for the frequency of earthquake aftershocks.

“We show that the public reception of news reports follow a similar statistic as earthquakes do,” the researchers conclude. “One might also think of a ‘Richter scale’ for media events.”

“I always think it’s interesting when people exploit the scale of online media to try to understand human behavior,” said Duncan Watts, a researcher at Yahoo! Research who describes himself as a “reformed physicist who has become a sociologist.”

But he notes that drawing mathematical analogies between unrelated phenomena doesn’t mean there’s any deeper connection. A lot of systems, including views on YouTube, activity on Facebook, number of tweets on Twitter, avalanches, forest fires, power outages and hurricanes all show frequency graphs similar to earthquakes.

“But they’re all generated by different processes,” Watts said. “To suggest that the same mechanism is at work here is kind of absurd. It sort of can’t be true.”

A couple of things are of note:

1. One of the advantages of the Internet as a medium is that people can fairly easily track these sorts of social phenomenon. The data is often in front of our eyes and once collected and put into a spreadsheet or data program is like any other dataset.

2. An interesting quote from the story: the “reformed physicist who has become a sociologist.” This pattern that looks similar to an earthquake is interesting. But sociologists would also want to know why this is the case and what factors affect the initial “wordquake” and subsequent aftershocks. (But it is interesting that the paper was developed by physicists: how many sociologists would look at this word frequency data and think of an earthquake pattern?)

2a. Just thinking about these word frequencies, how does this earthquake model differ from other options for looking at this sort of data? For example, researchers have used diffusion models to examine the spread of riots. Is a diffusion model better than an earthquake model for this phenomena?

3. Does this model offer any predictive power? That is, does it give us any insights into what words may set off “wordquakes” in the future?

0 thoughts on “Modeling “wordquakes”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s