Even though Google's Flu Trends, the big-data project that mines people's searches to uncover outbreaks of the flu, appears to have inflated estimates of flu cases, big data still hold promise, Northeastern University's David Lazer tells the Chronicle of Higher Education. Lazer and his colleagues wrote in Science this month that lessons can be gleaned from the errors made by the Google flu model.
"I would be quite distressed if this resulted in less resources being invested in Big Data," he says. Instead, he says this is "a good moment for Big Data, because it reflects the fact that there's some degree of maturing. Saying 'Big Data' isn't enough. You gotta be about doing Big Data right."
The problem, Lazer tells the Chronicle, was likely that the algorithm behind the tracker wasn't adjusted as Google tweaked its search algorithm to make it easier to search for health-related topics.
Similar issues have cropped up when using Twitter or other social media sites as data sources, the Chronicle notes.
Still, Nicholas Christakis, a social scientist and physician who directs Yale University's Human Nature Lab, says Big Data can help address a number of questions. Current issues, he tells the Chronicle, are "the birthing pains of that process."