The image shows the 50 phrases that generated the most buzz during the last three months of the presidential campaign. The vertical axis is number of web items; the horizontal axis is date.
This came from computer scientists at Cornell, but it seems like a harbinger of the type of text and meme analysis that will become common as the data and processing needed to do such work becomes easier to manage (see the work of Radian6, Crimson Hexagon and others).
There are so many interesting ways to look at data with a phrase-based approach. Love this ManyEyes "I need to" example with Twitter (warning: major time sink!): http://manyeyes.alphaworks.ibm.com/manyeyes/visualizations/twitter-i-need-to-___
Posted by: Perry Hewitt | March 13, 2009 at 10:52 AM
Whoa. That is cool.
What's really exciting to me is that it's not just a toy. You can really use this kind of stuff to draw conclusions about the type of stories that get traction or the types of things people want.
It means marketing, journalism and other content-related fields are more evidence-based, and less eminence based.
Posted by: Rick Burnes | March 14, 2009 at 05:32 PM