Glossary

Chartjunk

Chartjunk is Edward Tufte's term for visual elements in a graphic that do not represent data or help the viewer interpret data. Chartjunk dilutes the data-ink ratio, competes for attention, and often misleads the viewer.

Common examples:

  • 3D effects on bar charts (especially pseudo-perspective that distorts values)
  • Unnecessary gridlines darker than the data
  • Decorative backgrounds and watermarks
  • Clip art embedded in charts
  • Moiré patterns in shaded fills
  • Excessive colour for non-data purposes
  • Heavy borders and frames around plot areas
  • Redundant labels duplicating axis information
  • Inappropriate icons replacing simple bars or dots

Chartjunk is particularly harmful because it exploits pre-attentive processing: visually distinctive decorative elements capture attention that should be devoted to the data. A 3D exploded pie chart draws the eye to the decoration, not the numbers.

Tufte's preferred aesthetic — which he demonstrated in famous revisions of cluttered infographics — is almost extreme in its minimalism: remove every line, marker, and label that doesn't carry information. His redrawings typically delete 30–60% of the original ink while making the data more visible.

The term has entered general data visualisation vocabulary and is now used by practitioners who may not share Tufte's full aesthetic but accept the core principle: if it doesn't show data, seriously consider removing it.

Related terms: Data-Ink Ratio, Cognitive Load

Discussed in:

Also defined in: Textbook of Usability