# Tutorials

54 posts

The donut chart is not the most perceptually efficient chart. Although it might be a smidge better than its pied cousin. But if you’re making a donut chart and then animating it to show changes over time, I don’t think you’re chasing perceptual efficiency. Become a member for access to this — plus tutorials, courses, and guides.

Clusters of bubbles might not be the most visually precise way to show counts, but the elements can lend weight to the individuals that aggregates represent as a whole. The one-to-one ratio between element and count feels less abstract than a bar or a line. In this tutorial you learn how to create this one-to-one ratio. Become a member for access to this — plus tutorials, courses, and guides.

Maps are a good way to visualize spatial data, but to maintain geographic accuracy, there are often tradeoffs. For example, a choropleth map, which shades regions by a specified metric, maintains the same shape regardless of the data you color by. Larger regions always take up more space on the screen and smaller regions take up less space. Cartograms are a response to this challenge by sizing regions to match...

It’s typically straightforward to make and read a line chart. The position on the line represents a value, and the slope between points represents a rate of change. Usually a line chart that represents a single time series uses a solid color for the line. But while messing with a heatmap, which uses color as its primary visual encoding, I was curious what you could show if you introduced a...

In addition to a great format for visual jokes, Venn diagrams can clearly show overlap between sets. Sets, as in set theory, can be understood as objects belonging to one or more collections. The classic examples in mathematics are natural numbers, integers, rational and real numbers. But sets are everywhere. Become a member for access to this — plus tutorials, courses, and guides.

Many people were dismayed by a spiral chart that served as a header image for a New York Times Opinion piece. I thought it was fine. Others had other opinions. Disregarding whether or not it was the “best” way to visualize the data, clearly the more important question is how to make such a chart. Here’s how to make it in R. Read More

When you have many categories, use ridgelines to create an extremely compact visualization where you can easily identify major patterns and outliers. They are especially useful to display surges in mostly flat data series. Become a member for access to this — plus tutorials, courses, and guides.

Stacked area charts are a standard chart type that shows how a categorical distribution changes over time. There are packages that can do most of the work for you, but for a recent chart I found myself wanting more control over how the chart was drawn, which led me to drawing from scratch. Become a member for access to this — plus tutorials, courses, and guides.

ggplot2 provides sensible default settings for analysis, but when you make charts for a publication, you often need to match an existing style and shift focus to readability over exploration. Design around a message or results instead of leaving interpretation open-ended. Finally, you need to export your charts in the required file format with the correct dimensions and resolution. Become a member for access to this — plus tutorials, courses,...

Circle packing in visualization is a way to arrange circles in a fixed space so that none of the circles overlap, and if you were to increase the radius of any circle, it’d overlap with a neighbor. This can be a useful method to have in your toolbox to make various types of charts or to make existing charts more readable. Become a member for access to this — plus...