Friday, June 03, 2011

Mining patterns in search data with Google Correlate

May 31, 2011 by Editor From Kurzweil

Google Correlate has been launched on Google Labs. Users can now upload their own data series and retrieve a list of search terms based on popularity to see what best corresponds with a real world trend.
In the example below, official flu activity data were uploaded from the U.S. CDC over the last several years. The data reveal that people search for terms like [cold or flu] in a pattern similar to actual flu rates. Finding these correlated terms, a Google Flu Trends graph was built:
Google Flu
(Credit: Matt Mohebbi / Google)
Users can also enter search terms such as [ribosome] and find other terms with activity that correspond well over time:
Google Ribosome
(credit: Matt Mohebbi / Google)
white paper describes the methodology behind Google Correlate. How can Google help with the flu? Find out here.

No comments:

Post a Comment