The Internet is a key source of information for millions around the globe. Not surprising therefore that search engines, like Google are one of the first places people turn to for information on topics ranging from movies (Slum Dog Millionaire), overnight sensations (Susan Boyle), and threatening pandemics (swine flu).
The sheer numbers of people who use the Internet for search or Google’s dominant share of search engines is exciting at a mere wow level. More exciting is the fact that on-line search behavior of individuals litters the electronic highways with digital crumbs and telltale clues, leading to the obvious question – are they related to events in the real world and can they predict off-line behavior?
A team of researchers from Google and the CDC answer this question in a most compelling and topical way in the February issue of Nature. The authors were able to accurately model the outbreak of flu epidemics by tracking search engine query data. Their research tool, Google Flu Trends, has a warm intuitive basis to it – people are more likely to be searching for sunscreen during summer months and for flu remedies and prevention tips during the flu season.
The research team observed that some search queries tend to be popular exactly when the flu season is in full swing. A few examples:
- flu complications, syptoms
- cold/flu remedy
- antibiotic/antiviral medication
By observing and counting the frequency of these search queries the authors are able to accurately estimate how much flu is circulating in various regions of the United States.
The CDC also tracks influenza across the United States through their Influenza Sentinel Program, which relies on a network of approximately 2500 doctors who see 16 million patients each year. The doctors keep track of and report the percentage of their patients who have an influenza-like illness (ILI).
The CDC publishes national and regional data from these surveillance systems on a weekly basis, typically with a 1-2-week reporting lag. In an attempt to provide faster detection innovative surveillance systems using indirect signals of flu and flu-like activity like call center volume and sale of OTC drugs have been adopted recently.
While the CDC innovations are laudable, it is difficult to top the value of Google search queries as an early warning system for epidemics and pandemics. They can be counted automatically, quickly, results can be made available daily, and can be consistently published 1-2 weeks ahead of CDC ILI surveillance reports.
Early detection and rapid response are not just mantras for the commercial world. They are just as important in the worlds of institutional and government action. With over 90 million adults searching online for information about specific diseases or medical problems each year in USA alone, web search queries are an important line of defense for preventing and containing epidemics, before they become a pandemic.
It is in that spirit, that after conferring with US and Mexican health officials Google Flu Trends has created and released experimental flu activity estimates for Mexico based on aggregated search data. With the WHO raising the alert level concerning swine flue twice in the last three days, elevating it to one notch below a full-scale pandemic, all available data needs to be brought into play, regardless of whether its been validated against actual flu cases or not.
During difficult times like this, with the financial gloom still hanging heavy and swine flu threatening, it is encouraging to find social media being used to promote social well being, not just the goals of a few corporations with the right technology.
Web search logs, whether generated by Google or some other search engine, represent the collective intelligence of millions of Internet search users. We already have an example of how its being used intelligently for the early detection of influenza. Perhaps it can also be used to tease out early signals of a much-needed economic recovery!
What digital crumbs should we be looking for – search for air-fares, travel and holiday destinations, home prices, marriages and honeymoons?