I should have seen it coming. First it felled my boyfriend’s roommates, then my boyfriend, and then my roommate. Then, two weeks into the viral sensation sweeping the nation, I fell asleep with a sore throat, and woke up with a head full of mucus.
Luckily, it wasn’t the flu. But if it was, last week was also the first time I could have predicted when such a flu might strike my part of town, as it does during the peak flu months between October and April. That’s because, earlier this month, scientists at Columbia University’s Mailman School of Public Health uploaded a first-of-its-kind flu prediction model online.
In December, assistant professor of environmental health sciences Jeffrey Shaman told Co.Exist about the tool he and his colleagues had developed to predict the flu up to nine weeks in advance. Using data from Google Flu Trends and weekly CDC infection rates, the Columbia model was able to predict the exact timing of flu arrival accurately in 63% of the American cities it analyzed.
One day, Shaman suggested, the predictions might become so accurate that they’re eventually broadcast next to the weather on TV.
In the meantime, that model now exists on the good ‘ole Internet. It predicts some relief for Lincoln, Nebraska, which appears to be coming down from quite an illness, as does Wichita. Boston, on the other hand, looks like it’ll be experiencing an increase in flu cases over the next couple of weeks, as will New York City.
On the map above the predictor, you can check out CDC data for flu patient visits to the doctor’s office from the week prior. Next to the predictor, click on your state in the tree map to find out which cities will be most afflicted.