Early sketches from an aborted project looking at the email collection from the Enron case. Each layer is one person's email from the corpus. The layer is divided into portions using a Treemap representation based on how often each particular word is found in their text. Areas that are brighter show words that appear more often than the average across the entire mail collection.
The images began as a project for a print commission (a one or two page magazine spread) where I wanted to create the impression of being “inside” the information. Keep in mind that this is a sketch so it's not particularly useful, and the filtering isn't what it could be (e.g. some HTML tags are slipping in). Basically I went another direction after spending a few hours playing with the sketch.