I took some time out to analyse over 100,000 tweets about the Japanese tsunami yesterday. Although I haven't properly looked into the data yet, the levels of sentiment are incredibly interesting.
Below is a graph of the twitter sentiment across the majority of the day in the UK. Click on the image to view it at a larger scale.
For the record: I only started tracking the twitter sentiment after the tsunami had hit Japan. The following graph is the reaction to the event on Twitter.
Twitter sentiment during Japanese tsunami on the 11th of March 2011
Transmissions Received
Hip2b2
Interesting and impressive stuff. I have wondered and argued over the use of Twitter (my position was as a non-believer/user), but I now see great potential in using social media to not only communicate with people, but as a tool to assess zeitgeist (which is the really interesting stuff) in a way that transcends conventional polling. Of this potenial I am a believer and am looking to build some analysis myself, once I figure out how (any suggestions where to start?).
All this being said I am wondering why most of the people who are doing this sort of work on the net seem to be focused on generating visual representations as opposed to providing numeric views of the data. Is it the sexy factor, or something else? I also wonder why the underlying is never/seldom discussed. Take the Japanese Tsunami chart what word are being evaluated to define sentiment. Without these the whole exercise is somewhat dimensionless and doesn’t speak to me at least.
Regards
k