Overlapping dots in a crowded area quietly hide how many points are actually stacked there

What it is

A dot map places one mark per event or individual, with position carrying the geographic coordinate. Its systematic weakness is overplotting: when many points fall within the same patch of screen, they cover one another, and the density a reader perceives stops growing linearly with the true count. High-density regions are read as less busy than they are. This is not reader carelessness; it is the inevitable reading bias once the display channel saturates at the high end.

Why it happens

The saturation comes from the physical ceiling of screen rendering. Points have a finite pixel footprint, so only a limited number of marks can occupy any pixel area; points beyond that resolution collapse into the same mark. Visible density therefore has a ceiling: once the true count exceeds it, perceived density no longer tracks the count—the screen shows something close to min(actual points, displayable points). Readers estimate quantity from visible density, whether they count dots or judge busyness, so dense regions yield a systematic undercount. Sparse regions are unaffected and are in fact read most accurately, because individual dots stay countable. Transparency and jitter mitigate coverage without removing the underlying constraint: opacity makes overlaps darker rather than making more dots visible, and jitter moves the positions themselves.

Where it stops holding

The size of the bias depends on map resolution, point size, and opacity: zooming in raises the displayable point count per region and eases saturation; smaller, more transparent points push the saturation threshold up while making each dot harder to see. If the data is genuinely sparse—a few points per screen—the bias is negligible. Interactive maps let readers zoom until points separate, but the default full-extent view is still saturated, and judgments made without zooming already carry the bias.

Applying it

• Do not rely on a full-extent dot map for quantity comparison with dense data; either stop the default zoom level before dots visibly pile up or switch to an aggregated form.
• Use semi-transparent dots and accept that this only mitigates: the most crowded regions still cannot be read accurately.
• Verify by picking regions across a density range, asking readers to estimate counts from the map, and comparing against actual values; a systematic low bias in dense regions is direct evidence of saturation.

Related

• Same group: U12.03.2 Aggregation into clusters hides within-cluster distribution differences · U12.03.3 Cluster markers must encode how many points they contain · U12.03.4 Hexagonal binning or density surfaces can replace direct point overlay · U12.03.5 Zoom-dependent aggregation radii make the same data support different conclusions
• Nearby: U12.03.4 Hexagonal binning or density surfaces can replace direct point overlay · U12.02.1 Large-area regions gain disproportionate visual weight in choropleth maps
• Search terms: dot map · overplotting · point density