August 18, 2016 - Susan Rojo | Storytelling with Data

Brexit Insights from Data Visualization

The United Kingdom referendum to determine whether or not to stay within the European Union, commonly referred to as the Brexit, had people talking about one question: Stay or Leave? Let’s take the next step, and use visualization to let the data do the talking.

Talking dominated the runup to the vote, with a flood of campaigning from both those wanting to stay and those wanting to leave the EU; it even got a bit surreal right in the middle of the Thames. On June 23rd the vote delivered the answer. The same pundits who prognosticated the outcome — as many wrong as right, it seems — shifted the talk to  financial implications along with warnings about parallels to the coming vote in the US this November.

Brexit UKNow that a little time has passed, let’s go back to the data for insight into who actually voted to leave vs. remain. We used Arcadia Data to combine the polling results data with data about income and education across Scotland, Ireland, and England, to see what insights could be found.

For starters, the outcome we want to analyze is the national results, which had a modest spread between the two sides (51.9 % Leave vs 48.1 % Remain), but with a higher voter turnout (72 %) than in the previous election.

The UK Election Commission website has rich publicly available data on poll results. We found income data on the Office for National Statistics site and then researched the education data at the region and area level for Scotland, England and Wales, and Ireland, respectively. All said and done, this was not more than an hour of Googling around to get our hands on the data we needed.  The key was to sift out the above websites from the search results.  (It’s not always the top results to Google queries that are the most useful.)

We then loaded the data into Arcadia and modeled relationships between key fields in the data to begin to visualize the information we gathered. Since the polling data did not have geo-coding but was broken up by localities, we lined up the latitude and longitude equivalents for those so we could start looking at this geographically. We were then ready to quickly plot a map with circles that denote the percentage of voters who want to leave the EU. Each circle is a Local Authority (district) within the United Kingdom, and the Brexit Londoncolors indicate how strongly that particular district voted to leave the European Union. Interestingly, the rate of support within individual localities was not evenly distributed. In localities carried by ‘Remain’, the rate of support reached 96%. By contrast, localities carried by ‘Leave’ had a maximum support rate of 71.5%. We can observe that in some places where it won, the ‘Remain’ vote benefited from stronger local consensus than the opposing side. But that was not enough to win nationally.

Since the map layer is interactive, you can dig deeper to see further details. When we zoom in to the London area, for instance, you can clearly see the regional differences. London voters heavily supported staying in the EU (green dots) while the localities in surrounding areas voted strongly to leave (red).

After seeing an overview of how the vote played out across the UK, we were curious to take a closer look at the regional results and see what may have influenced the vote within these localities. We will share more on that in a future post.

Related Posts