Each dot = 1% of population Hover any dot Click a category to isolate Arrow keys navigate grid
A dot matrix chart represents a population as a uniform grid of discrete marks, each dot encoding exactly one unit — here, one percent. Categories are distinguished by color, with the grid filled sequentially. The perceptual mechanism is two-fold: counting (the viewer tallies dots in a category by scanning clusters) and area estimation (larger clusters read as larger proportions without requiring arithmetic). Because the total is always fixed — 100 dots — the viewer can compare proportions across all categories against a known denominator. This is the critical advantage over a bar chart: the shared visual budget makes part-to-whole reasoning immediate, not inferential.
The dataset is a population divided into five discrete, mutually exclusive categories that sum to 100%. The message is about proportional distribution — not trend, not correlation, not ranking. A pie chart serves the same purpose but activates angle estimation, the least accurate quantitative perceptual channel. A stacked bar works for single-series but becomes unwieldy for comparison. The dot matrix exploits position and grouping simultaneously: the visual system pre-attentively detects clusters of same-colored dots, making the largest categories visible before any label is read. The fixed 10x10 grid anchors the denominator visually in a way no pie chart can.
A standard pie chart fails in two ways. It encodes magnitude as arc angle — a channel where human perception is consistently less accurate than position judgments. And distinguishing five similarly-sized slices requires comparing non-adjacent arc lengths, introducing error. The dot matrix keeps all categories on the same spatial field. A treemap is the strongest alternative when the part-to-whole structure is hierarchical; for flat single-level proportions with a countable denominator the dot matrix is preferable because each unit is individually traceable — something treemaps sacrifice for space-efficiency.
FT Visual Vocabulary category: Part-to-whole — "Show how a single entity is divided into its component parts." Abela quadrant: Composition (how individual parts make up the whole, static). Tufte principle: every dot is data — the grid is the axis, no ink is decorative. The one design decision worth knowing: filling the grid row-by-row in shuffled order (rather than sorted by category) preserves the sense of a real population sample rather than a deliberately arranged bar chart — the apparent randomness is not a flaw, it is the encoding.