Wordle: Beautiful word clouds

Posted on August 27, 2008
Filed Under Play |

I am a sucker for beautiful and interesting data visualization techniques-  I’ve long been a fan of Edward Tufte, and his work, especially his Sparklines system for representing data in “intense, simple, word sized graphics”. Of course, there are other simple systems which have been made popular of late by the web, such as tag clouds.

While clouds might possibly be overused or abused in some designs, I think they’re just too fun and intutitive to pass up, and they can appear quite striking. 1 For example, Jonathan Feinberg created a cool web application called Wordle, that lets one visualize any arbitrary text.  One can either paste in text, or point the tool to a URL containing text to be processed.  The application is a Java applet, and it performs quite nicely, even with large amounts of text. The results are interactively configurable, and you end up with some pretty neat representations of word counts.

I took a couple of popular and lengthy texts, and had a play with Wordle- granted these are scaled down representations of my end results, but you get the picture, so to speak:

911report_small Wordle: Beautiful word clouds

Wordle cloud based upon the 911 Commission report

In case you cannot readily discern from the image itself, the above Wordle graphic represents the full text of the 9/11 Commission Report. I think it is very interesting and telling.  For example, numerous references occur to entities which one would expect, such as Bin Laden, CIA, etc. In a somewhat disturbing and telling revelation however, there are either no references, or not enough references to the word victim to generate anything here…

Moving right along, here’s a traditional religious text, ye olden King James’ very own edition of the Holy Bible- Whole lotta Shall’n goin’ on. This visualization is not surprising by any means at all:

bible_small Wordle: Beautiful word clouds

Wordle cloud based on the Holy Bible (KJV)

One of the projects I’m working on involves various kinds of data visualizations using the Python programming language and a slew of interesting datasets which I’ve been mining away at over the last few months.  I am hoping to draw much inspiration from folks like Tufte, and from Feinberg’s cool application, Wordle as well. Stay tuned, and keep your tinfoil hats crinkly, folks!

  1. Despite the whining about overuse of tag clouds by such cutting edge luminaries as Zeldman- really, just because something gets popular, it’s no longer going to be good? Woah- better take back fire and the wheel man, they are far too popular to be cool anymore! ;-P Seriously though; all things in moderation would suffice… [↩]

Comments

Leave a Reply