Software that tracks mood swings across the ‘blogosphere’ and pinpoints the events behind them could provide more insightful ways to search and analyse the web, researchers say.
The software, called MoodViews, was created by Gilad Mishne and colleagues at Amsterdam University, The Netherlands. It tracks about 10 million blogs hosted by the US service LiveJournal.
“I noticed that blog posts on LiveJournal have mood labels attached,” Mishne says. “We started to collect this information and noticed trends in different moods over time.”
About 250,000 new LiveJournal posts are created every day and roughly 150,000 of these include a label for one of hundreds of different moods. Moodviews keeps track of these labels and generates a graph, revealing emotions shifts across all LiveJournal blogs over time.
Moodviews reveals patterns that follow on weekly, monthly and even yearly cycles. For example, the label “drunk” becomes increasingly popular each weekend. The label “stressed” appears less during summer months and more towards the end of each year, perhaps because of end-of-year work deadlines or the stress of visiting in-laws.
On Valentine’s Day, there is spike in the numbers of bloggers who use the labels “loved” or “flirty”, but also an increase in the number who report feeling “lonely”.
The latest addition to Moodviews, a program called Moodsignals, tries to explain match these blogospheric mood swings to current events. It identifies emotional peaks by comparing recent label usage with records of previous use. When it finds a spike, the program picks out less commonly used words from relevant blog posts in an effort to identify the cause of the emotional change.
Moodsignals successfully determined that a peak in excitement and similar emotions over a weekend in July 2005 was related to the publication of a new Harry Potter book. “It found that words like ‘Harry’, ‘Potter’, ‘shop’ and ‘book’ were coming up more than usual in posts,” Mishne explains. “These terms were submitted to the news archive which showed the excitement was about the latest Harry Potter book.”
The tool may have commercial applications. For example, one investment banker is interested in using it to track consumer confidence in different products.
But a long-term goal is to create new ways to find information online and the researchers plan to release an emotionally aware search engine later in 2006, which will measure bloggers’ moods towards a particular word or topic.
Jill Walker, who researches online communication at the University of Bergen, in Norway, sees potential in this approach. “It would be a very interesting new way to navigate the internet,” she told New Scientist. “It is good to try and think of ways to access the kind of information that’s not easily machine-codable, but is out there on the web.”
Walker notes that can be difficult to divine feelings from online posts as people can choose to portray very different images of themselves online, compared to their real-world personas. But she says the information collected by Moodviews may still be useful. “We can find out a lot about how people choose to represent themselves online this way,” she notes. [article]
Very interesting, except if any software attempted to track MY mood swings, it would inevitably get ridiculously confused and probably crash.
Everything just has to be about me, doesn’t it….akh!