Washington program helps analyze a better Twitter through scienceon June 10, 2013 @ 6:33 am (Updated: 10:30 am - 6/10/13 )
At the Department of Energy's Pacific Northwest Laboratory in Richland, Wash., Court Corley is developing a program to analyze the billions of pieces of data that flood social media platforms every day. The computer scientist thinks that in a disaster situation, such as the Oklahoma tornadoes, tweets could provide crucial early information to help victims and first responders.
His program, SALSA, for Social Sensor Analytics, looks for trends and patterns in posts in more than 60 languages.
"We don't use keywords, we think the data should speak to us first, not the other way around," said Corley.
Weeding out the useless information is not an easy task considering there are 500 million tweets each day. So he samples and sorts the posts.
"What we do is look at the data in aggregate and the patterns, commonalities between words themselves tell us what's important," explained Corley.
Such as, where are power lines down, where are people trapped, where is help available? Corley is using a science-based approach to learn how to use social media in a more meaningful way.
Corley has a passion for public health and says Harvard and Children's Hospital of Boston have an app where people can report symptoms. He thinks tweets could provide an early indicator of disease, such as a recent flu outbreak in China where social media revealed the problem weeks before the government became aware.
In tomorrow's world?
"My hope is we'll all be living somewhere, over 100 and without disease," said Corley hopefully. By harnessing the power of the tweet, "to do things that never were possible before."
You might also like:
The 'White' Word
WWU's president defends his race-related comments to Dave Ross
Dori Monson says he loathes new taxes, but this one needs to happen
Please login below with your Facebook, Twitter, Google+ or Disqus account. Existing MyNorthwest account holders will need to create a new Disqus account or use one of the social logins provided below. Thank you.