The text is chopped into N chunks, and each "chapter" is plotted in a 2-dimensional space (connected by lines) along with the top X words in the text. You can see how the trajectory of the text moves through the space of words, emphasizing different themes at different stages of the work.

A bit more specifically, the text is chopped into N pieces, and each piece is treated as its own document. A word count matrix is created for this collection of "documents," and then my own implementation of the Nipals algorithm is used for a fast estimation of the principal components. Scores (for chunks) and loadings (for words) are standardarized so that the first two dimensions for both can be plotted on a single chart. The chunks are numbered and connected by lines.