Word Analysis of the Final Presidential Debate

Well the third and final debate is in the books. And this exhausting election will be over in less than three weeks! Like the first two debates, I thought it would be interesting to analyze the words used during the debate. Like my previous analyses, my goal is to understand which topics were addressed most frequently and how this contrasted between the two, how much each spoke overall, and how much time each candidate spent talking about their opponent. At times, I’ll also compare the data to my previous analyses.

Like the first two debates, I obtained the transcript from the Washington Post, then used a tool from WriteWords.org to parse out each word and its frequency of use. Once I had the frequencies of each word, I did some cleanup, filtering out pronouns, prepositions, conjunctions, and other types of common words, and removing any words which were used less than five times.

From the data, we can make a few observations:
  • In the first two debates, Trump spoke more overall words than Clinton (35% more in the first debate and 17% more in the second). However, in the third debate, Clinton spoke more words, using a total of 6,923, compared to Trump’s 6,681, a difference of about 4%.
  • Trump spoke a total of 1,037 unique words, while Clinton spoke 1,386, a 34% difference.
  • Like the first debate, both candidates spent a lot of time speaking about their opponent. Trump’s # 3 word (tied with “Great”) was “Hillary” (19 occurrences) and Clinton’s #2 word was “Donald” (28 occurrences). Like the second debate, Trump chose to use Secretary Clinton’s first name only (Clinton used this tactic in all three debates).
  • There was some of overlap in each candidate’s most commonly used words, particularly “People” and “Country”, but the rest of the words were quite different between the two candidates.
Here are each candidate’s top 10 most used words.

Word Cloud
Finally, like the first two debates, I thought that word clouds would be a good way to visualize this data set, so I created the following word clouds for each candidate.

And here’s a word cloud for Chris Wallace as well. Wallace, who I felt did a tremendous job of keeping things under control, spoke much more than the previous moderators. Not surprisingly, his top words were the candidates’ names—he said “Secretary” 49 times, “Clinton” 46 times, “Mr.” 44 times, and “Trump” 45 times.

Finally, I also created a bubble chart for each candidate, plus Chris Wallace.

If you have any questions or comments, please let me know. Next week, I’ll be posting my October “Nate Silver Challenge” update, so be sure to check it out. And, if you’d like to interact with the visualization, feel free to check it out on my Tableau Public page: https://public.tableau.com/profile/ken.flerlage#!/vizhome/Final2016PresidentialDebate/Bubble

Ken Flerlage, October 20, 2016

1 comment:

  1. I always liked this way of showing of data in the form of a presentation cuz it looks very understandable and clear. In my company I often have to create a lot of presentations of our products, that is why I use this service https://poweredtemplate.com/word-templates/index.html where I take a lot of templates.


Powered by Blogger.