Have you ever wondered if the Internet is having a good day, if it is happy or sad, or maybe feeling a little depressed? Well, probably not, but thanks to a team of information access researchers from Amsterdam University you might be able to find out.
MoodViews work by tapping into the mood level flags set by more than 10 million LiveJournal bloggers. By scanning the 150,000 or so daily postings that include one of the 132 different available mood flags, the MoodViews software collects the information and then looks for trends, Although there is a temptation to write this off as some kind of digital biorhythm, there is no doubting that emotional trends have emerged that make sense. So the Internet tends to be drunk at the weekend, feel loved on Valentine’s Day, was worried during Hurricane Katrina and positively scared after the 7/7 London bombings last year. The researchers do not see it as just a bit of fun however, but rather hope it will help them to develop new methods for searching, discovering and retrieving blogs.
You can keep an eye on how the Internet is feeling at any given time by checking out the MoodGrapher which updates and plots those mood levels every 10 minutes by showing the actual mood reported and the rate of change within any given hour. Then there’s the MoodTeller which uses natural language processing together with machine learning to guesstimate mood levels from the text of blog entries rather than those mood flags. Next on the launch agenda is MoodSpotter, due any time now, which will reveal those moods typically associated with any topic.