This is very important. How did your favorite candy rate?   The old FiveThirtyEight ran a head-to-head between Halloween candies years ago. I used the data for the above, which should be taken very seriously. Tags: candy, Halloween

Hey, it’s Nathan. This is the Process, the newsletter for FlowingData members where we talk about the finer details of analyzing data and chart-making. Every month, I collect tools and resources to make better charts and use data more effectively. This is the good stuff for October. Become a member for access to this — plus tutorials, courses, and...

For NYT Upshot, Ronda Kaysen and Ethan Singer compared current voter registration data against 2020 data to estimate how people moved away from politically balanced neighborhoods. Those who voted for a given party were much more likely to move to places that were similar politically. Particle-based flow charts are used to show the movements. They initially...

In NBA basketball, the home team used to almost always wear a white uniform and the away team would wear a non-white uniform that matched one of their team colors. Now it’s kind of all over the place. I assume it’s because Nike wanted to sell more jerseys. For The Pudding, Tyler Machado quantified the shift by manually counting uniform usage during...

For NYT Opinion, Gus Wezerek compared ad views per household: To figure out which party won this election’s ad war, Times Opinion dug into data on 3.7 million political ad airings on broadcast TV from Aug. 1 to Oct. 25. We combined those records, provided by AdImpact, with TV viewership estimates from Comscore to calculate how many households saw each...

By now you’ve seen or chosen to ignore forecasts that provide probabilities for a presidential candidate winning the election. Run enough elections with a given probability and the results converge. On the other hand, individual outcomes are a lot less certain. Read More