As a lifelong hypochondriac and a fan of the movie Outbreak, it was with great interest that I read Jared Keller’s new piece on Twitter and disease control on The Atlantic‘s blog. For the attention deficient among you, here’s the gist: because complaining makes us feel alive, people tweet about being sick. When lots of people in the same location tweet similar symptoms and/or diagnoses, it can be posited that a particular illness – the flu, say – is spreading rapidly in said location. Therefore, looking at data from Twitter is an easy way to manage disease outbreak. This is good.
Yes, there are flaws in the design, like the fact that still only a relatively small number of people in the world use Twitter. That number sinks significantly in places more at-risk for outbreaks, like Haiti, where there’s currently a cholera threat. The other problem is that there are no baseline stats: If five million people in New York tweet cold/flu symptoms, is it 1918 all over again, or is it just winter?
But for the most part, the use of Twitter as a data gathering device goes a long way towards finding some higher function for micro-blogging. A team at the University of Louisiana spent the last year connecting disease data on Twitter, and their findings had a 95% success rate in matching the CDC’s US flu predictions. Twitter data harvesters have an advantage over the CDC in that their data is updated instantaneously, whereas the CDC must sometimes wait days for the data to be gathered before it can be analyzed.
Data collection may also be the key to monetizing Twitter, a company whose formatting and design leaves little room for advertising, especially when companies can advertise for free – as many companies do – by simply tweeting their promotions. Remember, for example, when Rupert Murdock bought MySpace so he could use all the data collected on the site for market research? Unfortunately for Murdock it was right around the time that everyone stopped using MySpace. But Twitter is a market researchers dream, especially since almost all tweets (which the exception of the few Tweeters who keep their profiles private) are public information, available to anyone with a computer. The money will be in figuring out ways – like the Louisiana researchers did – of collecting relevant info, filtering it, analyzing it, and selling it to companies who could use that information.
So go start your Twitter-data company before someone else gets the idea. I’ll be home blogging and eagerly following Twitter so I know when the flu season’s over and it’s once again safe to leave my apartment.