Ian pointed me to a cool visualization project by Collins and colleagues at UOIT called Docuburst. Docuburst uses WordNet and visualizes the distribution of words in a WordNet tree starting from a node you select. Their code is made available.