Saturday, July 30, 2016

A Word Cloud of Trump and Clinton's Convention Speeches


It's official-- we can all stop tacking on the word "presumptive" when discussing the 2016 Presidential election. Hillary Clinton is officially the Democratic Party's nominee and Trump was deemed official last week.  In an unrelated note, millions of #NeverTrump-ers and Bernie supporters felt the last vestiges of hope crumble away. Although, to be fair, the writing has been on the wall for quite some time now. 



As is expected, both Clinton and Trump's speeches were well-received by their respective crowds and now the parties will begin the arduous task of trying to rally their members around the two least popular candidates amongst members of their own party in two decades (and, perhaps, the least popular candidates the country has seen in general).

I was interested in the speeches that the candidates gave at their nominating conventions, so I decided to make two word clouds that investigated what the candidates discussed. I acquired Trump's full speech from Politico and Clinton's from CNN. I then used R's tm package to make two separate corpuses (corpusi? I know it's not corpuscle) and the aptly titled wordcloud package to generate the cloud. After that it was just a matter of exporting the raw clouds and whipping up a pretty li'l PNG to make it a smidge more visually appealing. 
   

Astute observers will notice that there are a few odd-looking typos. What the heck is "presid" or "immigr?" This is what happened after I destemmed the words in their speeches. Destemming is when we analyse a set of words with similar core elements and reduce the linguistic accouterments to only reflect said elements. So if I was talking a lot about gallivanting over long distances like the masochist that I am, I would take words like "running" and reduce it down to "run." So "presid" reflects the frequency that Clinton discussed concepts relating to the presidency and "immigr" is when Trump talks about immigration or immigrants.

A couple of technical notes on the clouds: The clouds were constructed of words that occurred at least five times within each speech for a maximum of the top 75 words. The coloration does not reflect any sort of sentiment analysis--brighter words can be just as negative as darker words. The pallets were chosen only in a way to easily visually differentiate between the two.  Also: This does NOT allow us to compare how often Trump said a word compared to Clinton. The size of the words only relate to the frequency in which they occurred within each individual speech.


However, it is pretty interesting to see what both candidates discussed the most. The fact that both candidates use "will" pretty often seems fairly typical of political candidates. After all, you gotta focus on what YOU can do and on how the country is going to progress under your leadership. The fact that they both talk about the economy is also routine, which also explains the "millions" floating about there. These were generally used in the context of economic figures (e.g. millions of jobs created or lost)

A cursory look at Trump finds that he spent a lot more of his speech talking about issues that are threatening the country. "protect," "opponents," "kill," and "fail" shows up in there as well. Actually, "threaten" pops up in there verbatim, too. While some are taking ample advantage of this to point out how authoritarian and violent Trump is as a candidate, we cannot deny that these ideas resonate with a significant plurality of Americans with genuine concern over the country's well-being.

That's not to say that Trump spent the entire time discussing these sorts of issues. He also discussed taxes, the deficit ("trillion"), and the idea of making things "work" again. Now, whether things "worked" in the past or are indeed "broken" now is an empirical question-- one that definitely wouldn't be answered in Trump's favor-- but, again, it resonates with his demographic.

Clinton's speech was much more inclusive and policy oriented. There was discussion of doing things "together", "joining" together to accomplish goals (such as "health" and "pay"), "believing" in things in progress now and in the near-future. The term "good" explicitly appears too, almost as if to round off the fact that she was trying to make her speech as optimistic as possible.


Probably one of the most accurate uses of this meme ever to be seen on the internet.

That is, when she's not talking about Donald Trump. Which she does. Fairly often, too. On the one hand, considering her speech comes after Trump's, it makes perfect sense that she'd want to refute the opposition's rhetoric. (And it's not like Trump doesn't discuss Clinton; you can see her name hiding in his cloud, too). On the other, she does have a greater tendency to call him out by name--full name, actually-- so it wasn't as if she wasn't capable of turning on the attack. 

Some people are going to be inclined to see this criticism within Trump's speech and the optimism and inclusiveness in Clinton's speech as reflecting some sort of signal of candidate quality. ("Why would we want someone who's critical instead of inclusive?" is the sort of thing I've seen on social media this morning). However, we have to recognize that criticism of the status quo is an common tactic for parties that are not currently in power while optimism and inclusiveness is a hallmark of incumbents. After all, it'd be weird for a challenger to sing the praises of what's currently occurring. Almost as weird for someone currently in power to critique the work that they (presumably) spearheaded. Tone and attitude are not the best proxies for candidate quality. It's more of just a function of the times. 

It will be pretty interesting to see how Clinton and Trump's speeches evolve as the no-longer-presumptive candidates of their parties. Only time will tell if their words will be enough to convince the legions of disaffected voters to pick their poison...

Not that I'm projecting or anything. 



Peter Licari Student in Political Behavior and Elections

No comments:

Post a Comment