Categorical Data

Categorical data is data whose values fall into distinct named groups — like colour, region or product type — rather than along a measurable numeric continuum.

The defining feature is that the values are labels, not quantities. "Red," "blue" and "green" are categories; so are "small," "medium" and "large." You cannot meaningfully average them or measure the distance between them. The only natural operations are counting how many fall in each group and comparing those counts.

How categorical data appears in a chart

Because categories have no inherent numeric position, a chart places them at evenly spaced, arbitrary slots along an axis — usually the horizontal one. Each category gets its own bar, slice or column, and the value plotted is typically a count or total for that group. The order of categories is a design choice (alphabetical, by size, or a logical sequence), since unlike numbers there is no "correct" ordering for purely nominal groups.

A concrete example

Suppose a survey records favourite fruit and the tally is apple 40, banana 25, cherry 15, with 20 choosing something else, totalling 100 responses. "Apple," "banana," "cherry" and "other" are the categories; the counts are the measured values. A bar chart draws four bars whose heights are 40, 25, 15 and 20. The horizontal positions carry no numeric meaning — you could reorder the bars freely without changing what the data says.

Related terms

Categorical data contrasts with continuous data, which lives on a measurable scale. Each category typically forms one data series entry or group. Categories are the natural input for the bar chart guide; for continuous values binned into ranges, see the histogram guide.