Colors in Data Visualization: 200+ Years Ago and Today

Playfair’s charts, from the end of the 18th century to the beginning of the 19th century utilized single-color hand-engraved copperplate printing with the following hand-staining to incorporate a few additional colors.

Imagine the amount of work required to add a few colors by drawing a few lines and painting some areas on over 40 charts in 1,000 book copies (and that’s assuming you have already engraved those 40+ copper plates and printed the books).

Playfair’s Commercial and Political Atlas and Statistical Breviary, Cambridge University Press; First Edition (2005)

Now, it’s too easy to open a color picker and choose one of 16,777,216 colors. But data visualization is not photography. We need just one. Or two. Or three. Three out of 16 million. It’s too easy to click three times to get 3 hex codes. But choose wisely. And it’s not an easy task.

How to start? Resist data visualization tool attempts to turn your chart into a magic rainbow unicorn fairytale. It’s not an easy task either. Dragging and dropping a column into the “legend” field of a Power BI visual (with one of the predefined themes and default formatting settings) turns your visualization into a nuclear rainbow explosion. Stop it.

Make it black and grey on the white background. Later, add a few drops of color to highlight a few small, but important, things. Data visualization is about making a few small, but important, things visible, not about making everything glow and shine.