T.S. Eliot reads “The Love Song of J. Alfred Prufrock” while OpenWordle visualizes his words.
Dynamic OpenWordle
Here’s something neat you can do with OpenWordle:
It starts with a stream of words; in this case it’s Bonnie Tyler’s “Total Eclipse of the Heart” lyrics. (I chose it for the repetition and from remembering this interpretation.) The stream is tokenized but not reduced; this means duplicate words are included (and welcomed).
Normally during the layout, word information like position and color are discarded after the word is drawn. But if the information is saved, each word can grow as the data stream is processed. Older words are faded out to make way for new ones.
What do we learn about “Total Eclipse of the Heart”? She likes to turn now and then, but total eclipses are rare.
The Processing code that made this is included in the OpenWordle library download. The sketch is called “DynamicWordle” in the examples folder.
Here’s another one of Lady Gaga’s “Poker Face”:
Introducing OpenWordle
Since the Wordle layout algorithm was described in an InfoVis paper last October, I tried my hand at implementing it in Processing. Here is the result:
I made changes to the algorithm which resulted in more flexibility to add features like Unicode fonts and arbitrary text area confinement. I call this new project OpenWordle.
I’ve made the “prototype” code that created this available on the hosting site: http://code.google.com/p/openwordle. You can checkout the latest code from the repository or download a release. The code is covered by the GPLv3.
Why open source Wordle?
Wordle is great. As described in the paper “Participatory Visualization with Wordle,” many people have used it for many things, ranging from politics to literature to music lyrics. However, what the paper doesn’t cover are the unrealized possibilities of using Wordle because the code is closed source and thus not extensible. What about generating them for blogs ala tag clouds? What about collaboration between users entering data simultaneously? What about using it as a tool for real-time animation during, say, a presentation? I want people with application ideas to go ahead and do them. Open source code also allows bug fixes and feature requests (like Unicode font support and stemming) to be fulfilled.
As Jonathan Feinberg, creator of Wordle, states in his FAQ, “[The Wordle] code belongs to IBM, so it isn’t mine to share. I’m sorry that I can’t share it with you.” I have no reason to doubt his sincerity and I believe he would share the source if circumstances were different.
The Future
As OpenWordle only exists as a Processing sketch right now, there is still a lot to be done. I will try to create a Processing library, then a pure Java library. If you would like to help, email me with your Google account name (or Gmail address) and I’ll add you to the committers list. Also, I’d like to hear from anyone who builds a cool application with OpenWordle code.
A Complicated Definition
Back when I was a college freshman, there was an uproar caused by Avril Lavigne’s debut. She branded herself and her music as “punk,” an image that teenagers bought into. But the older punk veterans detested the label and decried her music as “pop” and herself as a fraud. For my part, since I had nothing against either genre, I thought it was amusing.
Recently, Manuel Lima posted an Information Visualization Manifesto on his Visual Complexity blog. It seems to have generated quite a bit of discussion, including a follow-up by Robert Kosara. In it he criticizes the overexposure of “frivolous” information art being labelled as information visualization. He then goes on to outline guidelines for what true information visualization should be.
Personally, I agree with some of his points and disagree with others. But this post is not about the details of the manifesto; it’s about the premise. I’ve seen what Manuel Lima is talking about: the data-limited pretty pictures shared on the web, created by artists and designers. Some are brilliant, some are dumb. I think the crux of what Mr. Lima is concerned about is the mislabeling of these art-oriented works as information visualization. Even with my liberal idea of what information visualization is, I agree that they are mislabeled.
But will having a strict definition of information visualization and guidelines for its practice fix the problem? No way. It’s an issue of semantics that will not be resolved with more semantics. I see it as a failure of the academic infovis world in educating the public about what we do. We are the ones writing esoteric research papers that are read by a relatively small circle of people. You can’t blame artists and designers for experimenting, using what they know, without following the infovis literature.
In the end, punk rock’s essence survived Avril Lavigne’s intrusion and her fans realized that she was indeed a pop singer. But the silver lining: How many young fans were introduced to the Ramones because of her?