Using the schedule and programme data, this prototype allows a user to locate programmes which are similar to one another and presents this in an easy to understand and interactive visual form.

At present, the algorithm for determining similarity is pretty rudimentary, simply comparing the genres of each programme. In the future, a more intelligent system could be put in place which could consider identifying comparable keywords in the synopses and also the target audience for the programmes, based on the time of day at which they are broadcast. Also, for now, it just searches BBC One, Two, Three and Four; but there is no reason that the user couldn’t have the option to choose which services to include.

http://www.nevett.org/lab/tv/similarity