Article summary
A new 'variance-in-time' method uses Google Trends (GT) to detect and forecast events by collecting multiple, overlapping GT samples over time. The algorithm leverages variation in both the mean and the variance of a search term to accommodate idiosyncrasies in the GT platform.