It's a Moody World
I subscribe to a news service called Future Brief which provides a daily email...
Normally 3 stories with just a paragraph on each, it is easily digested and some of their snippets are quite interesting. Today's included a link to an article in New Scientist (here) which describes a site launched in June 2005 that tracks mood swings across the 'blogosphere'.
Moodviews captures and provides analysis of posts on the US LiveJournal blog hosting service - a total of about 10 million blogs with about 250,000 new posts a day. About 60% of these posts use the 'mood tag' feature, where the poster can choose from 132 different moods. To the original Moodgrapher, which plots aggregate levels of use of each of the tags over time, they have added Moodteller, which does an analysis of the text of blog postings to determine mood.
Now they are adding Moodsignals, which detects words and phrases which are associated with a given mood in a given time interval, using statistical frequency comparisons and burstiness models. It looks at unusual peaks in the levels of mood annotations, and then tries to explain the peaks found by analyzing the language used by bloggers. In looking for explanations, Moodsignals searches news archives.
And, finally, they also provide a Moodsticker giving the latest information on the most popular mood, updated every 10 minutes.
Fascinating, but I'm finding it difficult to see applications of this outside market research and PR. I wonder if there are any mashup applications of this in retail?
...to the busy professional who wants to remain aware of trends in a rapidly changing world, yet not be buried under a mountain of information
Normally 3 stories with just a paragraph on each, it is easily digested and some of their snippets are quite interesting. Today's included a link to an article in New Scientist (here) which describes a site launched in June 2005 that tracks mood swings across the 'blogosphere'.
Moodviews captures and provides analysis of posts on the US LiveJournal blog hosting service - a total of about 10 million blogs with about 250,000 new posts a day. About 60% of these posts use the 'mood tag' feature, where the poster can choose from 132 different moods. To the original Moodgrapher, which plots aggregate levels of use of each of the tags over time, they have added Moodteller, which does an analysis of the text of blog postings to determine mood.
Now they are adding Moodsignals, which detects words and phrases which are associated with a given mood in a given time interval, using statistical frequency comparisons and burstiness models. It looks at unusual peaks in the levels of mood annotations, and then tries to explain the peaks found by analyzing the language used by bloggers. In looking for explanations, Moodsignals searches news archives.
And, finally, they also provide a Moodsticker giving the latest information on the most popular mood, updated every 10 minutes.
Fascinating, but I'm finding it difficult to see applications of this outside market research and PR. I wonder if there are any mashup applications of this in retail?
posted by John | Friday, April 21, 2006
0 Comments:
Post a Comment
<< Home