Where Do the Candidates Visit Most Often?
Ken Flerlage
Tableau Evangelist & Consultant at Moxy Analytics | Tableau Visionary Hall of Fame | Tableau Forums Ambassador | One Half of the Flerlage Twins
As you may have noticed from many of my posts, I’m a bit of a politics junkie. I read everything I can about the election—good, bad, and (mostly) ugly—I watch lots of election coverage on TV, and I listen to a variety of politics podcasts. My favorite of these podcasts, by far, is the NPR Politics Podcast. The show is hosted by campaign reporter Sam Sanders, and regulars include White House correspondent Tamara Keith, political editor Domenico Montanaro, political reporter Danielle Kurtzleben, and editor/correspondent Ron Elving, among others. They do an incredible job of making sense of this crazy election and I would highly recommend you check out their show. You can find more details here: https://www.npr.org/podcasts/510310/npr-politics-podcast
As I was listening to the show on Thursday, they answered the following listener question:
I would like to know what factors go into where the candidates visit. I live in eastern North Carolina, not far from where Trump asked African Americans what we have to lose and the area where he delivered this speech is a rural area, comprised mostly of farmland. He delivered another speech in another rural area tonight, Wednesday. I live about 40 minutes away in a city with two Marine bases and, with the exception of Jill Biden, no one has made any campaign stops here. It seems to me this would be a prime campaigning ground. What determines where candidates make campaign stops? Is it the population or demographics?
I thought this was a very interesting question and I really wanted to get a better understanding of where they go and, perhaps, why. Fortunately, I found a website called “2016 Travel Tracker” (https://traveltracker.nationaljournal.com) which compiles information about the candidates’ campaign stops and provides some maps and other analytics. While the site has some great information, I wanted to explore the data even further, so I pulled the data off the site, starting with September 1, and did some of my own analysis (Note: The following visualizations show September 1 through November 1).
My first question was a simple one: Who makes the most campaign stops? I created the following unit chart in Tableau to help answer this question.
Donald Trump has made far more campaign stops than Hillary Clinton—his 85 stops are 81% more than Clinton’s 47. In fact, both Tim Kaine and Mike Pence have also made more stops than Clinton, with 57 and 51, respectively.
The next question I had was, in general, where are they spending their time?
Not surprisingly, most of their time is spent in the so-called “battleground states”, particularly those battlegrounds with the highest number of electoral votes, Florida, North Carolina, Ohio, and Pennsylvania. From this, I began to wonder how closely the number of stops and the number of electoral votes are correlated. So I created a scatter plot comparing the two.
While there appeared to be some correlation between the two, there are a lot of outliers. But when you examine these outliers, you see states like California, Texas, New York, Illinois, and the District of Columbia, all of which are safely in the hands of either Clinton or Trump (there has been some question about Texas, but I personally see little reason for concern for Republicans). So, perhaps we should remove these states and only show the battleground states (Note: I'm using my list of battleground states, which I updated in my September "Nate Silver Election Challenge Update" https://www.dhirubhai.net/pulse/nate-silver-election-challenge-september-update-ken-flerlage)
This changes the story quite a bit and there starts to be a clearer correlation between the number of stops and the electoral votes. Florida, with 29 electoral votes was visited 43 times, followed by Pennsylvania (20 electoral votes) with 31 visits and Ohio (18 electoral votes) with 30 visits. There are still a couple of outliers, but both of these—New Hampshire and Michigan—have seemed largely out of reach for Trump most of this election cycle.
Finally, I wanted to get back to the NPR Politics Podcast listener question, which theorized that the Republican team is spending more of its time in somewhat rural areas. To understand this better, I mapped each candidate’s visits, overlaid on top of a background map showing population density. First, we’ll start with a map of all four candidates together, which will give us an overall feel for which areas are being visited most often.
Now, we’ll look at each of the four candidates individually.
Donald Trump:
Mike Pence:
Hillary Clinton:
Tim Kaine:
Note: The darker the background color, the higher the population density. The darker the dot, the more recently the candidate visited. And the larger the dot, the more visits the candidate made.
The maps do seem to lend some credence to the listener’s theory. If you take Ohio, for example, both Clinton and Trump have visited the larger population centers of Cleveland and Cincinnati, but Trump also visited smaller towns like Wilmington, Canton, Springfield, and Geneva, while Clinton has kept her visits to the large population centers of Akron and Columbus (in addition to Cleveland and Cincinnati).
So, there you have it. If you would like to interact with the visualization discussed in this post, you can find it on Tableau Public. The visualization allows you to analyze based on additional information and also includes Gary Johnson and his running mate, Bill Weld. https://public.tableau.com/profile/ken.flerlage#!/vizhome/2016CampaignEvents/Events. If you have any questions, comments, or additional insights, please let me know.
Ken Flerlage, November 1, 2016
Website: www.kenflerlage.com
Tableau Public: https://public.tableau.com/profile/ken.flerlage#!/
In need of restructuring your accounts ? Getting some analysis and reporting done on your business ?
8 年Ken, Nice visualisations. And good questions asked. But the weakness of data visualisation as an analysis tool shows giving possible correlations and explanations only.