Vocabulary TreeMap
Treemaps are used to show hierarchical information through nested rectangles. The size of each rectangle represents some quantity - numbers of times that words occur in the collection here. Rectangle positions show the relationships between different levels of the hierarchy and their colours relates to numeric values. Here the hierarchy shows words, within (occurring in) books, within (written by) authors.
Our interactive treemap allows us to explore the 100 most important words in each of the publications. These words are represented by rectangles sized according to the number of times they occur in a document. The largest rectangles occur to the top left showing the authors who use the most words, the books with most words for each author and the most popular words in each book at different levels of the hierarchy. Publications with more words cover a larger area - as do authors with larger total word counts.
So, Dickens has the most words under consideration here and can be seen to be the author of 8 books. Of these, Martin Chuzzlewit has the most words, more than twice as many as Oliver Twist, but fewer than Tolstoy's War and Peace. Shakespeare uses more words than Tolstoy, but in far fewer documents.
The colours show the importance of the words in each publication through a metric known as the TF-IDF. Words shaded in the darker reds are the most important according to this measure. So, a larger box that is also dark indicates that an important word in a publication is also frequent.
To use the visualization move the mouse over the words to see the author and publication, as well as the TF-IDF of that word and its frequency.
- Drag up or down with right mouse button to zoom in or out.
- Drag with the left mouse button to pan.
- R restores to the original view
- C changes to a different ordering. Here books are ordered from the top left by publication date rather than number of words and authors are ordered by date of birth - allowing us to explore the way that language varies between authors and over time.
Advanced Functionality for Further Exploration
- H - toggle highlighting of the selected word across the publications.
- L - hide / show labels.
- : - highlight words beginning with particular letter combinations (press again to clear)
- I - summary information panel on / off
- K - shows a legend
- hold down the A - highlight author level
- hold down the B - highlight book level
- hold down the W - highlight word level