Basic sentiment analysis is easy, any fool can do it let’s be honest.
You start off with a list of positive words and a list of negative words, figure out on the content you want sentiment on how the positives and negatives faired.
That gives you a basic score. If want to try it for yourself then you can read this walkthrough and that’ll get you started.
Sentiment goes away beyond a Twitter stream though quite a few companies have made a good living from doing it (and I certainly don’t want to take that away from any of them at all). It goes way beyond that.
The XFactor is a good example. The sentiment is carried further, from judge influenced comments. Judges will protect and defend their own acts obviously and you can hear it in their comments to others not in their team. Is anyone measuring that content? Worth keeping in mind when there are betting odds at stake.
Press sentiment, depending on the news source, carries varying weight depending on the news source. Reuters will sentiment will carry more weight that The Sun. It’s all sentiment and it all needs measuring.
A lot of eggs have been put in one basket as sentiment analysis goes, mainly due to the fact that we’re willing to accept that rolling update services like twitter are the be all and end all of public opinion. They’re not. Keep in mind that you’re getting any decent sense out of 40% of the tweets consumed. It’s a difficult act to perfect, there are some good companies out there doing but not many.
With the exception of Google who can cover the real quantity of content that needs measured?
I’ll say it one more time just in case you didn’t read it properly.
Keep that in mind when you’re using someone elses data in your startup. 🙂