One Way To Spot A Partisan Gerrymander

Category: Best visualization (small and large newsrooms)

Country/area: United States

Organisation: FiveThirtyEight

Organisation size: Small

Publication date: 15/03/2019

Credit: Ella Koeze, William T. Adler

Project description:

This project is a scroll-based explainer of “partisan bias” a specific measure of how skewed an electoral map might be. Partisan bias was mentioned in the Supreme Court case on partisan gerrymandering in North Carolina and this piece first explains how the metric works using congressional elections data from the past 30 years, and then uses real historical political maps from three states to talk about some recent instances of gerrymandering and corrective maps that have been put in place as solutions. 

Impact reached:

The project was well recieverd and cited by gerrymandering experts (such as creator of the efficiency gap Nicholas Stephanopoulos) as a clear explainer of a widely used metric in gerrymandering. Gerrymandered maps are notoriously hard to define (just ask the Supreme Court) but this project gives readers a clear explanation of one way to go about it. 

Techniques/technologies used:

The project was built primarily with D3.js and Stickybits.js for the scroll-based animation. Because this was a complicated thing to explain, we felt that going step-by-step using scroll was the way to go. The analysis of elections data was done in Python and R. We also built a statistical model to account for unopposed elections.

What was the hardest part of this project?

There were three hardest parts to this project: the first was collecting all the congressional vote data going back to 1992. We had some of this data already in our databases, but much of it we had to find and collect ourselves. The second was building the statistical model to account for unopposed races. We considered excluding unopposed races from the analysis, but realized it was integral to the integrity of the project, so we did the extra work to account for them. The last most-difficult part was structuring the step-by-step words and animation for clarity and sense. It took much editing and user testing to come to the right sequence.

What can others learn from this project?

Firstly, we hope they can learn about partisan bias and some of the history of gerrymandering in America over the last three decades. I also think this is an instructive piece in how scroll and animation can help clearly explain a complicated concept. 

Project links: