Tracking tweets by dialectS

There are many ways to find the location of people on the internet. Some people are located through technology — but some people give away their location by their word choice. Internet speech is more regional than we'd previously thought.

Many people have bemoaned modern technology, saying that it leads the world into a gray, uniform hegemonic wasteland, in which everyone talks, acts, and thinks the same. When they do so to you, you can shoot back that vibrant differences in speech are still alive and well, especially when it comes to Twitter. (That will really get them.) Analysis of tweets show strong regional dialects expressed in 140 characters or less.

Word choice is a big divider. People persist in calling soda 'pop' even though it makes them sound like they're 1950s ragamuffins with catchers mitts stuffed in their back pockets. People in other regions name-check, calling it Coke. Even something as subtle as spelling is separated by geography. Northern Californians tweet cool as 'koo' while Southern ones use 'coo'. Sadly, "pimp" is also heavily skewing towards Northern California.

Although preliminary, the analysis indicates that we take slang, word choice, and spelling more from the people physically around us than people on the other side of the country. Our mass assimilation isn't complete yet.

Via Science Daily and MIT.