This is one of the videos we've shown at BlackHat Barcelona. It shows how well the new Named Entity Recognition (NER) transforms perform.
We start with a phrase ('uranium enrichment'), then look at PDF documents containing the phrase, run the OpenCalias and Alchemy API for NER on it, ending up with a graph that shows person's names, locations and companies - all in a Edge Weighted View (EWV) graph with referring URLs. Finally we compare the top locations found with what's listed on Wikipedia. The video is rather large (150 MB) but shows Maltego v3 in HD and should be worth watching.
The NER transforms are currently available on both the commercial and community Transform Application Servers (TAS) - you just need to follow the transform discovery wizard to get them.
Have fun and if you have a lot of fun - take photos...;)
PS: yes - some parts of the video have been sped up to avoid boredom and a larger file.