Making Twitter trends meaningful
Clive Thompson has a good post where he complains about the banality of Twitters trending topics for 2009:
Instead of showing trending topics from the entire "Twitter population", Thompson suggests to filter specific subgroups and analyze their tweets for common words. That's certainly a very interesting idea; unfortunately, none of the tools he mentions in his post produce satisfying results. With the new Twitter list functionality, it would be great (and, I guess, pretty easy) to have trending topics for every Twitter list (I would, for example, love to have the trending topics for the Sandboxers Twitter list). If the analysis is broken down to specific lists, the results over a longer period of time and especially the shifts (like appearance and disappearance of a specific term) could potentially be very interesting.
The problem here is the problem with all mainstream, middle-of-the-road subjects: They’re not going to be surprising, and information that isn’t surprising often isn’t useful or interesting either. (The way I see it, mainstream topics mostly useful in social bonding: Heavily-trod subject matter is great when you need to make pleasant chatter with strangers. “Crazy weather we’re having, eh?”) It’s not that the concept of “trending topics” is itself useless. It’s that sampling the entire population of Twitter is often pointless, because it’s too broad.
Instead of showing trending topics from the entire "Twitter population", Thompson suggests to filter specific subgroups and analyze their tweets for common words. That's certainly a very interesting idea; unfortunately, none of the tools he mentions in his post produce satisfying results. With the new Twitter list functionality, it would be great (and, I guess, pretty easy) to have trending topics for every Twitter list (I would, for example, love to have the trending topics for the Sandboxers Twitter list). If the analysis is broken down to specific lists, the results over a longer period of time and especially the shifts (like appearance and disappearance of a specific term) could potentially be very interesting.
