One of the things I love about that place known fondly as "the Interwebs" is the way it allows researchers to graph things that should never be graphed. For example, have you ever wondered exactly how excited people really were about the release of the most recent Harry Potter book? Thanks to MoodGrapher, an application created by three Dutch information theorists, you'll discover that reported feelings of excitement were up 130 percent on the day millions of copies of Harry Potter and the Half-Blood Prince flooded into stores.
MoodGrapher works by collecting information from the "mood" tags associated with millions of entries on popular diary site LiveJournal.com. Every time you write something on LiveJournal, you have the option to tag your post with a mood from a pop-up menu that includes everything from "bored" to "drunk" (for those feeling eccentric or curmudgeonly, there is also a "fill in your mood" option). After aggregating these moods over time, the Dutch info geeks were able to see clear patterns. Drunkenness spikes on Fridays and Saturdays, for example. Frustration plummeted on New Year's Day (but loneliness was on the rise). You can search for moods over time yourself if you visit the MoodGrapher at ilps.science.uva.nl/MoodViews/Moodgrapher.
The idea of tracking the moods of an entire global population sounds like something from a movie about a dystopian future in which humanity's computer overlords monitor everyone's feelings so they can dope us up or feed us rock and roll accordingly. And that's not far from the truth. The MoodGrapher's creators published a paper earlier this month suggesting that their tool could be used to predict the success of a given movie by measuring the warmth of people's feelings about it before release.
Some might argue that this is a consumer-centric development, in which our feelings are taken into account before new pop culture is thrust upon us. But in point of fact, measuring people's moods about something before it comes out merely reveals how much buzz has been generated by advertising campaigns. Thus, the MoodGrapher's results simply reflect how much money has already been blown on getting LiveJournal weenies amped up for the latest Franz Ferdinand album or M. Night Shyamalan's stupid new movie. There are some exceptions to this, certainly. But you're unlikely to see mass upticks in excitement for a new thing — whether it's Windows Vista or Joss Whedon's Wonder Woman movie — unless it's already being hyped to death.
What's truly interesting about the MoodGrapher isn't its marketability but its use as a diagnostic tool to measure how much events in the news affect people's emotions. In a paper called "Why Are They Excited?" the MoodGrapher team explain how they figure out what's causing unusual spikes in the mass mood. First they use a simple algorithm to search for massive mood upturns or downturns in a given period of time (usually a day or an hour). Then they search the journal entries of everyone who has reported the popular mood, looking for words or phrases that are used repeatedly. Once they've gotten five or six recurring words (like book and prince, for example), they search a news database to find out whether the words are turning up there too.
Using this methodology, the MoodGraphers found that a peak in "excitement" on July 16, 2005, was heavily correlated with the use of words like book and read and Potter. Similarly, they found that a peak in the mood "worried" during late August 2005 was associated with uses of the words hurricane, gas, and Katrina. Quick searches of those words against their news database revealed what you'd expect: They were ripped from the headlines.