Thursday, March 29, 2007
More Random Copying
The press is starting to pick up the Random Copying story. To the right is a graph of the rate of occurrences over time since the story first "broke" yesterday at about noon. Let's see if the popularity of the story takes the shape of a random variant. It would be great to be able to do this on the fly - generating histograms of events from news.google.com based on queries. Google "trends" does this to some degree, but it uses search terms to build its data - not frequency of reporting in the news media. Thus its measuring something different - the rate at which people go to the web to look for a term -- not the rate at which it appears on the web. Building an application to do this shouldn't too hard given the automated way in which news.google.com search are generated. One would just need to parse for time and title of each entry (assuming each is really related to the topic of interest). Hmmmm.. I'll have to cruft something up. Maybe later.