Data visualization classics

John Rauser's Velocity Amsterdam 2016 talk on how human vision reads graphs, built on William Cleveland's research into graphical perception. It explains which visual encodings people judge accurately and how that should shape the charts you make.

Nathaniel Smith and Stéfan van der Walt present the thinking behind Matplotlib's new default colormap, viridis, at SciPy 2015. They show why the old jet rainbow map distorts data, explain perceptual uniformity and color spaces, and cover how the new map holds up under color blindness and in greyscale.

In this clip from the BBC's The Joy of Stats, Hans Rosling animates health and wealth data for 200 countries over 200 years in four minutes, using moving bubble charts. He narrates the animation like a sports commentator.

Hans Rosling's TED talk, in which he uses Gapminder's animated charts of income and life expectancy to show that the old split between rich and poor countries no longer describes the world. It is a data talk in which the presentation of the data is part of the argument.

Hans Rosling's TED talk on global population growth, using boxes to represent people and show how the world's numbers are likely to level off. It argues from data, with child mortality and family size at the center of the story.

Anna Rosling Rönnlund of Gapminder presents Dollar Street, a project built from photos of homes at many income levels taken in dozens of countries. Sorting the photos of everyday things like beds, stoves and toys by income shows that how people live depends more on income than on nationality.

Vox argues in this short video that a chart's y-axis does not always have to start at zero. It explains when a truncated axis is reasonable and when it misleads.

A Vox explainer on how charts of the coronavirus pandemic can mislead readers. It looks at how choices in scale, data and presentation change what a chart appears to say.

Matt Parker of Stand-up Maths explains the Datasaurus Dozen, sets of points with the same summary statistics but very different shapes when plotted. He talks with the people behind the datasets about why plotting your data matters.

Physicist Angela Collier explains, in an extended rant, why she thinks violin plots should not be used to show data. It is a critique of one chart type and what it fails to show.

Neil Halloran's data-driven documentary on the human cost of World War II, using animated charts and maps to show how many died, where, and how that compares with other conflicts. It lets the viewer feel the scale of the numbers by pacing the visuals alongside narration.

Isao Hashimoto's animation plots every nuclear explosion from 1945 to 1998 on a world map, flashing each test as it happens and keeping a running count for each country. It is a short piece that turns a data set into a time-lapse.

Václav Volhejn explains how he built maps that show time instead of space, warping the map to reflect travel time. He walks through the design problems of making such a map.

A visualization from the American Museum of Natural History that flies outward from Earth through the solar system and galaxy to the edge of the observable universe, using astronomical survey data.

Mike Bostock, creator of D3, speaks at the OpenVis conference on design as a search problem. As the title says, he frames design as a search through a space of alternatives rather than a single inspired idea.

Kanit Wongsuphasawat, Dominik Moritz and Arvind Satyanarayan present Vega-Lite, a high-level grammar for building interactive graphics. Charts and their interactions are described as compact declarative specifications that compile down to the lower-level Vega.

Bret Victor demonstrates a prototype tool for drawing data-driven visualizations directly, without writing code in the usual way. The talk argues for creative tools that let people work on graphics and data with direct manipulation.

Math professor Chris Staecker demonstrates a 1915 nomogram, a printed graphical device for solving the quadratic formula without calculating. It is one of his short videos on old calculating tools.

An animated explainer of the 2008 credit crisis, told with simple drawn graphics and narration. It walks through how the crisis built up and spread.

The adumb channel builds a graph of Wikipedia articles and their links, then explores the clusters and structure that appear. It is a data visualization project explained as a story of what the network showed.