We’ve moved over to our new server, and the site is back up and running.

Please feel free to jump right in – it should feel nice and zippy. We’re still working on making sure all the old links and embedded visualizations behave properly, so don’t be too concerned. We’ll have all of that working over the next day or so.

If you see anything unusual, feel free to contact us via the usual channels.

One of the most popular visualizations on Many Eyes is the tag cloud. It’s easy to see the appeal: a tag cloud gives a quick, strong sense of the gist of a text. A good example is the image below, which shows Martin Luther King’s famous “I have a dream” speech.

But a tag cloud scrambles the flow of the text, reducing it to a collection of disconnected words. What if you want to see more structure, such as the poetic repetition of King’s oratory? What if you want to find recurrent themes, or explore the context in which terms are used? Our new Word Tree visualization is designed to let you analyze text in its original context, seeing the ways that words and phrases are used and repeated throughout
the work. Here’s an example (click for the fully interactive version):

A word tree starts by finding all the occurrences of a search term (in this case, “I”), along with the phrases that follow that term (”have a dream,” “must say,” etc.) It then groups these phrases so that, for instance, all the “have a dream” phrases are placed together. These groups form a tree that shows you some of the structure of the text, with thematic strains emerging. You can choose to view the tree of phrases that follow your search term, or the phrases that lead up to it.

In addition to typing into the search field, you can navigate the word tree by clicking on its branches. A click will narrow the view to an individual branch. Control-clicking a word is like following a hyperlink and will re-center the tree. This is a fast way to navigate a text - and our favorite way of using the visualization!

The word tree is a new visualization technique still in the experimental stage, and we’re curious to hear your feedback.