Maths boffins of the University of Bristol's Intelligent Systems Lab have found another great use for Twitter – to work out if it's raining. They can also use the social network to monitor the prevalence of flu, according to freshly published research.
Using "state of the art" algorithms, Professor Nello Cristianini and Vasileios Lampos compared rain reports with location-tagged tweets culled from Twitter that mentioned keywords relating to rain. They did the same with NHS records and tweets about flu, concluding that, yes, what people waffle on about via social networks really does reflect what's going on in the real world.
They reckon this is going to be pretty useful. The abstract of the research states:
We present a general methodology for inferring the occurrence and magnitude of an event or phenomenon by exploring the rich amount of unstructured textual information on the social part of the web.
The new-found ability to discover the bleeding' obvious by looking at Twitter will be helpful for pinpointing things, the researchers believe; they also announced their intention to keep up the research: "Future work could be focused on improving various subtasks in the methodology, enabling researchers to become ever more expert at pinpointing situations, such as a flu outbreak or electoral voting intentions."
They acknowledge that Twitter's 200 million users do not represent the general population – we're looking at 2.85 percent of the world's seven billion human inhabitants – but maintain that this study indicates that Twitter can be used to track an event.
It follows on from their work the previous year on, er, Twitter and Flu.
Nowcasting Events from the Social Web with Statistical Learning by Vasileios Lampos and Nello Cristianini was published in ACM Transactions on Intelligent Systems and Technology (TIST), September 2011. An abstract can be found on the Bristol University Website. ®