It’s politics month here at Tableau Public! We are preparing for Election Day in the US (November 4th people, mark your calendars!) with our #VizTheVote theme, and so we thought we would write some blog posts about visualizing voting data and election results. I wanted to explore one of the most controversial recent elections in my blog post, seeing how a Tableau visualization could shed light on what now, 14 years later, still rages on as a fiercely contested topic: the 2000 Presidential Election in the state of Florida. I’ll first present my viz, showing just how close the race in Florida was between George Bush and Al Gore, before briefly going over how I made it, and then analyzing it.
Finding data on elections is quite easy; a simple Google search will yield a plethora of different options for data exploration. It is important to know what type of data you are looking for, however, to expedite the data acquisition process. I knew that I wanted to represent the 2000 Florida election on a map of the state showing in which respective geographic areas Bush and Gore won, so I needed to find data with voting results broken down on a county level. I was able to do so by downloading their “Voting Results” file.
Now that I had my data, I could connect it to Tableau and start making my viz. The first thing I did was to try and create the map, thinking that it would be the foundation of my dashboard. As I started to create a filled map of Florida broken down by county, however, I was immediately struck with a problem, as 36 of Florida’s counties were unrecognized by Tableau’s mapping function. To remedy this, I used a simple trick that can help your mapping usage: I clicked on the “36 unknowns” message in the bottom right of the map, clicked on “edit locations” and then, in the “geographic roles” pane, I set the Country to United States and State to Florida, thus directing Tableau to plot all my county names only within the State of Florida, and thus eliminating possible country-wide redundancy. As seen in the before and after screenshots below, with this simple trick, all the ambiguous counties were matched up to their proper geographical location.
With my mechanical issue sorted out, it was fairly simple to create the map you see on my dashboard: I created a calculated field for Bush % and Gore % by dividing my Bush/Gore votes fields by the sum of all votes in each county and then coloring the map on a red/blue diverging palette based on the percentage of county residents that voted for each candidate.
My other two sheets in the dashboard were also easy to create: The sheet showing total votes in the three categories of the election – balloted, absentee, and total – are just bar charts that I colored with each candidate’s party colors. The absentee votes sheet is a stacked bar chart showing how many absentee votes the two candidates received in each category. A writeup of how to create stacked bar charts can be found here.
Disclaimer: I majored in Political Science so I love political data. Now that that’s out of the way, let’s get into some analysis of the dashboard, always my favorite part of vizzing!
How important was this election? The winner would take the White House and win the Presidency.
How close was this election? Before the absentees were counted – absentee ballots that represented only 0.04% of the total voting pool in Florida – Gore won the state election by absurdly narrow margin of 202 votes. Then the absentee ballots were counted (not without considerable controversy) and Bush pulled ahead by a (very relatively) robust 547 votes, winning Florida and with it the Presidency.
A closer look at the absentee results show that one county, Duvall, with its large military population, disproportionally helped leapfrog Bush to victory. It is notable that the results of the balloted election had Bush holding a 57% - 41% edge over Gore in Duvall county, but in the absentee pool, Bush’s edge climbed to 67% - 33%.
Cue the trumpets calling shenanigans, but I am afraid that a full revisit (not revizit!) of the recount controversy is, most unfortunately, just a little bit outside the scope of this blog post.