A recent article on the BBC news site describes how the police used twitter to identify new targets for rioters but that very quickly they were swamped with the number of tweets and comments that were being made.
In addition, they found it difficult to sift out which comments were serious, and which were not. If you are sitting reading the tweets as they are made then this will of course be very difficult – when the volume reaches a certain level it will be virtually impossible for anyone to be able to make sense of it.
However the analysis presented previously in our blog shows how an automated contextual analysis tool can indeed be used to make sense of the underlying data. We were able to analyse the data on an hourly basis, and simply ask what locations are coming up in context with the word ‘next’?
The analysis presented in our whitepaper shows clearly that the locations most talked about were indeed attacked by rioters later on. Our analysis picked up the locations discussed in the BBC article as well as other locations such as Brixton, Croydon, Walthamstow, and so on. In almost all cases there was a time gap of a good few hours between the location being discussed on twitter and the rioting and looting activity actually taking place.
This shows again how the use of contextual analysis on social media can be used to gather information, and indeed how powerful this approach can be.