acm-header
Sign In

Communications of the ACM

ACM TechNews

The Emerging Pitfalls of Nowcasting With Big Data


View as: Print Mobile App Share:
Searches for the word hangover rise substantially on a Saturday, peak on a Sunday, and drop off on a Monday. That is similar to the pattern of searches for the word vodka, albeit lagging by a day.

Google chief economist Hal Varian warns that search query data must be treated with care and caution.

Credit: Technology Review

Search query data is very powerful but must be treated with some care and caution, according to Google chief economist Hal Varian at a European Bank workshop.

For example, University College London researchers have studied Google Flu Trends data in which Google uses the number of flu-related searches to nowcast the incidence of flu in different parts of the world at any particular time. Although there are examples in which Google accurately estimated the number of flu cases, there also are examples in which Google Trends significantly overestimated the actual number of flu cases, the researchers learned.

The reason for the disparity is that some users are making flu-related searches because they are suffering from flu-related symptoms, while other users are searching because of other factors, such as strong media interest in the flu.

The researchers theorize the pattern of independent searches over time will differ substantially from social searches. The researchers also emphasize independent searches should increase rapidly as flu sweeps through a population and decline slowly as the disease dies out, while social searches are more symmetrical.

From Technology Review
View Full Article

 

Abstracts Copyright © 2014 Information Inc., Bethesda, Maryland, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account