I was recently asked by a colleague how useful I consider Twitter to be. I thought it to be just another social networking platform, but it can be used to generate some interesting information.
First of all, the website of Summize Labs is well worth a visit. One can enter a topic in the Summize Labs search engine to find up-to-the-second "tweets" about that topic, then automatically analyze the attitudes expressed in the "tweets". (Tweets come from Twitter - in case you wondering). As an example, the overall sentiment on obama is "swell". This could prove to be a very powerful tool for political scientists, marketers and all kinds of researchers.
Whats more, the Summize Labs blog, desribes how to use a local operator in twitter search, enabling one to search within tweets near a location. For example, you can find what people are saying about Obama near:oregon while he campaigns there. The easiest way to find this operator is to use their advanced search page --- there you will see an input box with "near this place." Just fill it in with your city, state, zip, etc. along with what you are looking for. This combines with GIS with novel attitudinal data --- very powerful indeed! (For those interested in GIS/GPS, Summize Labs use the Google Maps API to interpret (or "geocode") a free-text location as an actual place that can be put on a map).
These methods have considerable advantages over Twitter's public timeline, TweetScan and Quotably. But the question remains - is Twitter a very select sample? I don't use it and I don't know anyone else who does...
3 comments:
I had never come across this before.It looks interesting.I submitted "Liam" and its search concluded that (and I quote) "The overall sentiment on this topic is great".I think this may reflect that stripy shirt Liam has been wearing of late.
they are different stripy shirts - got two more for my birthday
i have some doubts about the usefulness of this:
examples of searches and results below. i started hoping to get a feel for how some of the disciplines where faring and then i was simply trying (unsuccesfully) to find something it didnt like.
"economics": swell
"psychology": great
"neuroscience": great
"murder": great
"simon cowell": great
"adolf hitler": swell
"torture": swell
"terrorism": swell
"animal cruelty": so-so
you kind of get the impression that its hard to get a bad score. although given that "swell" is worse than "great" it is noted that economic ranks below murder, neuroscience, psychology and simon cowell, while being just on a par with torture, adolf hitler and terrorism. Its generating more positive twitters than animal cruelty though so that's nice.
Post a Comment