This new algorithm from researchers at Columbia University’s Mailman School of Public Health combines data from Google Flu Trends and weekly reporting from the CDC’s surveillance network. Apparently, the addition of the CDC data dramatically improved predictions over just using Google Flu Trends.

How accurate and predictive would such models be if we could combine data from Google, Facebook, and Twitter with real-time results from a large network of hospital microbiology labs? Hospital micro labs generate a ton of useful data each day, yet it sits siloed within each hospitals’ walls.