What's interesting about the algorithm is the way it connects the "fading" technology of printed news with the onset of digital media: A major resource feeding her algorithms is an archive of The New York Times, along with Twitter feeds and Wikipedia entries. Because Radinsky can now identify cause-and-effect patterns with this system, she can alert us to possible disaster, political events, and even disease outbreak. “If a storm comes two years after a drought, a few weeks [after the storm] the probability of a cholera outbreak is huge, especially in countries with low GDP and low concentration of clean water,” she explains to Fast Company.
So, just how accurate is she? About 70% to 90%. Her algorithm predicted the cholera epidemic in Cuba (the first in decades) as well as the riots that sparked the Arab Spring. While it may seem that a bit of common sense and a lot of research could allow many scientists to forsee something like a cholera outbreak, Fast Company notes the real innovation is in Radinsky's automation of it: "Getting a computer to do it, and to analyze accurately the massive amounts of electronic data present on the web, is another matter."
What's important to remember about Radinksy's algorithms is that they suggest probability, rather than certainty. When she began fiddling around with Google Trends in 2007, Radinsky quickly realized she could predict what people would search for based on breaking news stories. Perhaps the best illustration of her technique is in her findings on the Arab Spring riots. Though her software successfully forecast the riots, it also predicted fall of the Sudanese government, which didn't happen.
Now, Radinsky has formed her own start-up, SalesPredict, a sales and marketing predictions organiations that dedicates a portion of its research to medical and humanitarian endeavors in collaboration with SparkBeyond. Most recently, her team predicted a cholera epidemic in Zimbabwe that could break before 2014. It's the hope that such warnings can be used to help us better prepare for troubling times, in turn making them more manageable. (Fast Company)