Tools

Which Chart Should I Use?

Every dataset has one chart that tells its story clearly — and a dozen that muddle it. Answer a few plain-English questions about what you're showing and the kind of data you have, and get the right chart: a live thumbnail, when to use it, and the classic mistake to sidestep.

A starting point, not a rule. The best chart depends on your data and your audience. When in doubt, show the data — a scatter or box plot reveals more than a bar of means — and always label your axes and start bar counts at zero. For the theory, see Visualizing Data; for putting them in a report honestly, Tables & Figures That Don't Lie.

The whole map on one screen

Prefer to scan? The same map as a table: goal and data type in, chart out.

Your goalData typeReach for
See one variable's shapeOne quantitativeHistogram
Compare center & spread across groupsQuantitative × categoricalBox plot(s)
Compare one value per categoryQuantitative × categorical (few)Bar chart
Rank many categories / long labelsQuantitative × categorical (many)Horizontal bar chart
Compare across two categorical splitsQuantitative × two categoricalGrouped / stacked bar
See how two numbers relateTwo quantitativeScatter plot
Summarize or predict a relationshipTwo quantitativeScatter + trend line
Show change over timeQuantitative × timeLine chart
Show association of two categoriesTwo categoricalHeatmap / mosaic
Show parts of a wholeCategorical proportionsStacked bar (not a pie)
Let readers read exact valuesAnyTable