Like millions of Americans, I eagerly watched the first Presidential debate, wondering what the next four years would bring. As I watched and listened, my mind started wandering to visualizations which could slice and dice the words of the verbal debate and help bring some greater insight into the debate.
If you are on a mobile device, you won’t be able to run the visualizations as they require Java Applets. If you are on a device which supports Java Applets, clicking the image will take you to that visualization. They are interactive, so you can tailor them for the issues which interest you.
Just the Facts Jack
No hype, no opinion, just computers hammering away at the available data.
I started by downloading the transcript from CNN, putting the data into a text component inside of Dex. From there, I filtered empty lines and parsed the conversation in order to isolate the debate into a tabular format which looked like:
From there, I saved a copy of this data, created an in-memory database table so that I could isolate things which were said by Barack Obama versus those statements which were said by Mitt Romney.
Unfortunately, Dex (at this time), lacks any meaningful form of text analysis. I considered using the Text Mining module through the Dex integration, but dismissed it once I realized that deep text analytics in R would not be mastered overnight. Instead, I turned to a very user friendly visualization tool called Many Eyes.
Many Eyes has some very powerful visualizations. I start with the Tag Cloud.
Word Trees are perhaps my favorite visualization of words. They preserve the phrases and the context of the words. In the images, I search on the world “health”, showing relevant word trees for both Barack Obama and Mitt Romney. I encourage readers to click the links and search on other relevant terms such as: taxes, economy, budget, jobs, education, etc…
I hope that you enjoyed these visualizations and that they prove useful in making an informed decision.