EU referendum: The result in maps and charts

EU referendum: The result in maps and charts

Organisation: BBC News Visual and Data journalism team (United Kingdom)

Publication Date: 04/06/2017

Size of team/newsroom:large

Description

The biggest UK story last year was the EU Referendum. The BBC covered the referendum results live on the night with a purpose built election results system but the role of the data team was to analyse those results and create fast turnaround data visualisations for the following day for the news website and social media. As the results continued to come in we quickly published and updated maps of vote share for Leave and Remain and turnout using an interactive map we had already prepared. We also integrated a postcode search into the page which would return the result by local authority. Then we did simple charts looking at the top areas for remain and a regional stacked mekko chart showing how Leave had won. We had gathered as much potentially relevant background council-level data as possible before results day, including on age, gender, ethnicity, educational attainment, social class, and country of birth, and prepared a workflow to combine this data with the results data as soon as it was available. As the results came in, we used linear regression modelling to find the strongest correlations between each of these variables and support for Leave and Remain. We found that the strongest positive correlations with support for Leave were the proportion of old people and the proportion of people identifying as English, and the strongest negative correlation was with the proportion of university graduates, and communicated these links as clearly and simply as we could, using three maps.

What makes this project innovative? What was its impact?

There was an immense appetite for information on the morning of the EU referendum results. Our data vis and analysis was published very fast - with the bones of it there as early risers were coming to the news website - and was built on across the day. Each hour or so we published a new chart or visualisation which updated the page and was also posted on the main BBC News accounts on Twitter and Facebook. This is not a traditional way of publishing data analysis but we found it to be very effective - and used the same model in the US elections. The lack of an official exit poll meant that it was difficult to explore the demographics of the result in a meaningful way. But anticipating this we used the council level population data we had prepared and statistical modelling to give original demographic insights into the result, bring real value to our analysis. The page got more than 4.5 million page views on the day of publication and was one of the top stories on the site across that day and the next.

Technologies used for this project:

We used Python, Jupyter, Pandas, Matplotlib, QGIS as well as well as Adobe Illustrator and JavaScript.
Follow this project
Wait
Comment

Comments (0)

You have to be connected to contribute

You have to be connected to follow

Leave this project and no longer be informed about this project

By joining this project, you will be informed by email when an update or a new contribution is posted on the website.

Thank you for your active participation !
Best,

The GEN Community Team