Muddy helps drive the BBC meta services engine, which enables relevant content to be dynamically retrieved and also suggested to journalists.

An example is the BBC Topic pages, which pull together relevant information around a concept, in this case Barack Obama. This information is retrieved from different ‘silos’ around the BBC and to be able to do this the BBC needs to be able to 'ground' concepts, that is to make sure they are referring to the same thing. Muddy helps to provide these grounded concepts as part of the content production process. A Topic page like Barack Obama will have quite a bit of editorial control but less prominent concepts could be more dynamic.

This project, built by Rattle and commissioned by the BBC Research & Developement department, used Muddy to analyse the subtitles of TV programmes.

By identifying the people, places and organisations mentioned on TV, the project was able to use additional information from DBpedia and Freebase to generate meaningful statistics about TV channels - such as the average age of the people mentioned, or the relative prominence of Northern vs Southern cities.

The project was launched as a prototype, and is available at http://channelography.rattlecentral.com.