Election 2016: Analyzing Media Coverage of the Candidates on Twitter
This post is featured as part of Election Month here on Tableau Public.
Rarely, if ever, has there been a shortage of scrutiny around journalistic bias in election reporting and the 2016 cycle is no different. This time around, mainstream media outlets came under fire for over-covering the Trump campaign.
To get to the bottom of the balance of distribution of candidate coverage in the media, I decided to turn to the platform nearest and dearest to this election’s heart, Twitter.
In order to get the data I needed to do an analysis on candidate coverage by media accounts, I used the ever-awesome app IFTTT. I automatically pulled tweets from CBS, CNN, NBC, MSNBC, and Fox News from Aug. 1 to Sept. 1, and amassed a HUGE number of tweets.
Before I could start my analysis, I need to clean up the tweets, I used Rapidminer, a data science tool I recently came across, to get rid of punctuation, links and unimportant stop words. Then classified them as relating to Hillary Clinton, Donald Trump or both candidates. Finally, I wanted to assess the sentiment of these tweets, which I did using a text processing extension in Rapidminer paired with the SentiWordNet dictionary. Once this process was finished, each tweet was labeled as positive, negative, or neutral.
Now that I had all the additional metadata I needed and my collection was formatted correctly, it was time to start vizzing. I started off by asking my initial question: As far as volume was concerned, which candidate was mentioned the most? Then I got a little more granular and looked at which candidate each of the six outlets talked about the most, and finally, how that coverage changed over time.
Once I answered these high-level questions, I addressed the big questions: What were the sentiments of these tweets? Were certain outlets’ tweets about one candidate more positive or negative than the other? Who had the most objective tweets?
As you can see in the viz below, Trump did indeed get a lot more coverage across the board.
Visualizing these tweets got me to take a step back and really think about the how news is served to up and reminded me that I need to step out of my information bubble every once in a while and take a look at what's circulating outside of my personal stream.