Project description

Over the past year, FiveThirtyEight has published everything from large, multi-page election forecasts to an interactive game that pitted readers against each other to teach them about game theory and trade wars. We created (live updating!) beautiful illustrations of in-game win probablities with area charts during the World Cup and explained how complicated gerrymandering metrics work with a scroll based explainer. And we kept ourselves accountable by evaluating every prediction we’ve made since 2008 and teaching readers how to evaluate forecasts.

What makes this project innovative?

FiveThirtyEight sets the bar for using statistical models and probablistic thinking in our journalism and our visualizations. We don't shy away from using visuals to explain complex methodological information such as the inputs to our elections model or how different measures of gerrymandering work. We also use live data to power up-to-the-minute forecasts for events such as soccer games and elections. We don't want to just describe news or events, but rather help readers understand context for why things happen and when they are significant. We approach journalism from an almost scientific standpoint, and we use data visualizations to not only show a result, but also to explain how we got there.

What was the impact of your project? How did you measure it?

Our work is part of the daily lives of many of our reader who want to check the latest poll numbers, election odds or sports probabilities. During live sporting events or elections nights our graphics serve as a second screen experience for hundreds of thousands of readers. We also build projects that use reader input directly and teach people about themselves. Our scientific version of the classic personality quiz was taken (from start to finish) by more than 850 thousand people. And our "How to Win a Trade War" project asked readers to engage in a game theory duel more than 1.5 million times.

Source and methodology

We use a wide array of data sources to power our work. We also create a lot of our own data such as probabilities and ratings. In all our work, we strive to be as transparent as possible and provide clear descriptions of our methodologies, either within a project or in a standalone methodology article.

Technologies Used

Our team uses a diverse set of technologies, including but not limited to: Python, R, Ruby, JavaScript, D3, SQL, node.js . The exact mix varies for each project.

Project members

Jay Boice, Aaron Bycoffe, Rachael Dottle, Ritchie King, Ella Koeze, Dhrumil Mehta, Andrei Sheinkman, Gus Wezerek, Julia Wolfe


Additional links


Click Follow to keep up with the evolution of this project:
you will receive a notification anytime the project leader updates the project page.