Recent posts

Playful with Words

We’ve been focused on words in the last few months, nouns mainly. We have two approaches to words: top down and bottom up, Muddy and Wordr. Our top down version is Muddy. Muddy extracts entities (known words) from any document / site and says which of those entities are ‘notable’ (i.e. deemed to be significant), [...]

Playing and pitching

This past month has seen us playing with BBC programmes and closed-captions data using Muddy as well as finishing off a couple of projects like this University of Sheffield project.  The BBC programmes work has been particularly interesting. You can glimpse some of the outputs from the @channelography twitter feed. We’ve also been pitching this [...]