Twitter Idea Of The Day
I just read Clive Owen's piece in on wired.com describing the rise of search engines focused on real-time trend monitoring, as opposed to indexing based on authority. It's good, short, and I recommend it.
Building on ideas I had a while back, it provoked an idea for a web service that would allow a group sponsor to register Twitter feeds (or, for that matter, any kind of feed) from members of the group, do a word-frequency analysis on those feeds (with appropriate filters of course), and then display snapshots (perhaps with a word cloud) of popularity, and trend analysis (fastest-rising, fastest-falling). You could also have specialized content variants: most popular URLs, most popular tags. Clicking through from any particular word (or url or tag) you could do a network analysis: which member of the group first mentioned the item, who re-tweeted him or her, either with attribution or without.
The builder of a service like this would construct it as a platform that would allow group sponsors to set up individual accounts with one or more groups, and it would allow these sponsors to aggregate groups up or drill down from an aggregate cross-group view down to individual ones, perhaps with some comparative analysis -- "show me the relative popularity of any given word / content item across my groups", for example.
Twitter already has trending topics, as do others, but the lack of grouping for folks relevant to me makes it (judging by the typical results) barely interesting and generally useless to me. There are visual views of news, like Newsmap, but they pre-filter content by focusing on published news stories.
An additional layer of sophistication based on semantic analysis technology like, say, Crimson Hexagon's, would translate individual key words into broader categories of meaning from all this, so you could, at a glance, in what ways and proportions your group members were feeling about different things: "Well, it's Monday morning, and 2/3 of my users are feeling 'anxious' about work, while 1/3 are feeling 'inspired' on vacation."
As for making money, buzz-tracking services are already bought by / licensed by / subscribed to by a number of organizations. I could see a two-stage model here where group sponsors who aggregate and process their members' feeds could then re-syndicate fine-grained analysis of those feeds to media and other organizations to whom that aggregated view would be useful. "What are alumni of university X / readers of magazine Y focused on right now?" The high-level cuts would be free, perhaps used to drive traffic.