Twittervision: From Cool, To Tool

Following on the recent post I wrote about Twitter, it's occurred to me that location is a major feature of many of the use cases I envisioned.  Today I came across Twittervision, an interesting Google Maps mashup described here. Picture scenarios that filter Twittervision into logical groups (pre-defined groups of people watching a webcast, or people linking and reacting to news, for example).  Now picture further parsing the Twitter posts for the occurrence of keywords signaling reactions, a la We Feel Fine, and then mapping those occurrences to a heat map overlay on the Twittervision Google Map.  What emerges is an "evolving geospatialized map of emotional reaction to events".  Surely this has to be useful to some news organization?