The word cloud is an interesting idea. Maybe normalizing it to find out which words were relatively frequent would be a thing