How Do You Feel about Sentiment Analysis?
Glancing at my predictions for 2011, I wish I could get a mulligan on the reputation one, which is a like a wish that I keep hoping will come true.
It won’t, at least not in 2011.
If I had a do-over, I’d predict something awesome for sentiment analysis. After all, we’ve been talking about Twitter, Facebook and sentiment, e.g. twistori, for a while.
Two stories this week point to a big year:
Fflick rates movies based on tweets about the movie. Twitter has become a key piece of movie marketing, as studios recognize and try to harness its ability to predict box office performance. No surprises here really. Going to see a movie is a core “what are you doing?” activity, as is discussing the movie with friends.
Everyone understands the power of social recommendations, and movies are an area where they carry a lot of weight. Google, said to be focusing more on social, is smart to buy up this algorithm, which can be easily rolled into its overall search capabilities. Google already uses the frequency of search keywords, e.g. the Google Flu Trends, to plot and model behavior.
Predicting the stock market is a bit more involved, but the relationship between mood (as interpreted by certain words appearing in tweets) and the DJIA produced a startling correlation. Researchers found that words indicating calm preceded rises in the market by three or four days.
There’s a lot of science in there, but that’s pretty amazing.
Anyway, as more people tweet and post status updates, the more data exist to map emotions to real world events. This is really cool stuff.
Thoughts?
Possibly Related Posts
- Check out Oracle Events
- The Really Smart Phone
- Facebook Knows When You Need a Hug
- Feeling Lucky?
- Twitter Break: Unexpected Use Cases, Search
-
uvox
-
http://theappslab.com Jake
-
http://sethgrimes.com Seth Grimes
-
http://theappslab.com Jake



