This is just a quick update to shed some light on a tool that I'm creating for visualising Twitter sentiment in real-time using Web technologies. I've already worked on Twitter sentiment analysis using historic data, so moving into the real-time world seemed like the next logical step.
Over the past weekend I put together a first attempt at creating a real-time visualisation of sentiment on Twitter over the past hour, using nothing but JavaScript and HTML5. Each segment on the graph is a single minute of data. Right now this is only a measurement of about 1-4% of tweets sent within the UK. However, this is still a considerable number that provides us with interesting results.
Note: This video is playing 3 times faster than normal. You'd get a little bored otherwise.
Transmissions Received
Rio
This is really neat! I wonder if there’s a way to localize it by a certain location and then run the queries as such?
Jake
This is great !
Are going to write up a tutorial on how to create this ?
It would be great to see what methods you have used to analysis the data
Paul M. Watson
Nice work Rob, shows how powerful the client-side web is becoming. What was used for the actual sentiment analysis? Twitter’s positive/negative search?