Leading up to the inception of Google Flu Trends in 2009, experts at Google realized they could forecast flu activity by tracking the frequency of when millions of people Google specific flu-related terms.
Unlike the data from the CDC, which at times can be potentially outdated, the Google Flu Trends tracker uses aggregated real-time data in order to monitor and forecast potential threats of flu infection.
While the application seems in par with the CDC’s reporting of flu occurrences, as seen in the tool’s historic charts, the Google Flu Trends tracker does appear far superior in reporting time – again, providing users with real-time daily data in lieu of week old reporting.
However it seems Google’s Flu Trends tracker has suffered a few technical sniffles along the way, according to a critique of the online application in the journal Science. The report, “The Parable of Google Flu: Traps in Big Data Analysis,” states that while the data provided by the online tool is quicker it is not always better. Nor is it as accurate as it’s been touted.
The study was funded, in part, by a grant from the National Science Foundation. Ryan Kennedy, David Lazer, Alex Vespignani (Northeastern University) and Gary King (Harvard University) – contributed to the research.
Together they examined Google’s data-aggregating tool Google Flu Trends (GFT), which was designed to provide real-time global monitoring of flu cases. “Google Flu Trends is an amazing piece of engineering and a very useful tool, but it also illustrates where ‘big data’ analysis can go wrong,” said Kennedy, who is a University of Houston political science professor.
Even with modifications to the GFT, the tool created to improve response to flu outbreaks has overestimated peak flu cases especially in the US over the past two years. The team found that GFT overestimated the prevalence of flu in the 2012-2013 season, as well as the actual levels of flu in 2011-2012, by more than 50 percent. Additionally, from August 2011 to September 2013, GFT over-predicted the prevalence of flu in 100 out of 108 weeks.
“It missed by a huge amount last year and actually, it turns out, it’s been missing by a fair amount for several years,” says David Lazer, a professor of political science and computer science at Northeastern University, in response to the findings.
According to NPR, Lazer finds the old school method of painstakingly collecting data across the country and forwarding to the US Centers for Disease Control and Prevention (CDC) still more superior in comparison to the Google Flu Trends site, even with the delay.
He states in Science Daily, “You could just have used old CDC data — two or three weeks old — and have projected forward, and done a better job than Google Flu Trends.” And while it seems Lazer and his colleagues may find the flu tracking technology unreliable, the professor does think the core idea behind the online tool is a terrific one.
Still, the team believes the application requires a few more tweaks. They suggest Google work in collaboration with a few outside scientists in order to improve the Google Flu Trends forecasting ability.
Research aside, there are several helpful smartphone apps available to consumers that essentially do the same thing: Sickweather, Flucast, and a free app called a Flu Near You.
[Photo Credit: William Brawley]