Scroll to zoom · Drag to pan · Hover dots for details · 1 dot = 1 branch
A Dot Map (Point Distribution Map) encodes the precise geographic location of each discrete observation as an equally-sized dot plotted over a geographic base layer. Unlike a choropleth map — which aggregates data by administrative boundary and encodes totals as fill colour — the dot map preserves the exact spatial coordinates of every individual entity. The primary perceptual mechanism is preattentive density detection: clusters of dots register as regions of high concentration before conscious counting begins, while sparse areas read as low concentration. Two variants exist. In a one-to-one map, each dot is a single real-world object. In a one-to-many map, each dot represents a count (e.g., 1 dot = 100 residents). This is a one-to-one map: every dot is one confirmed library branch at a real coordinate.
This map plots confirmed public library branch locations across the contiguous United States. Each walnut-coloured dot represents one physical branch location with a recorded name, city, and state. The spatial pattern that emerges is striking: library density is highest across the Northeast corridor from Boston to Washington D.C., and across the Great Lakes region including Ohio, Michigan, and Illinois — reflecting historical settlement density and robust civic funding traditions. The interior West — Nevada, Wyoming, Montana, and the Dakotas — shows sparse coverage, a direct consequence of low population density and vast geographic area per resident. Use the region filter buttons to isolate and compare density across the four Census Bureau regions. Scroll to zoom into any dense cluster to resolve individual branches that appear merged at full extent.
A Dot Map encodes the geographic location of discrete events or objects as equally-sized points plotted over a geographic base layer. It exploits the viewer's perceptual ability to detect visual density — clusters of points register immediately as regions of high concentration, sparse areas as low concentration. This is a preattentive task: the pattern emerges before conscious counting begins.
Two variants exist. In a one-to-one map, each dot represents a single real-world object — one library, one hospital, one crime incident. In a one-to-many map, each dot represents a unit count (e.g., 1 dot = 100 residents). This implementation is one-to-one: every dot is a confirmed branch location with a name and coordinates.
The message is about spatial distribution and clustering — where things are, and whether they concentrate. No aggregation is applied: each point is a real facility at a real coordinate. Aggregating to county or state would mask the intra-regional variation that is the actual story (a state with one dense urban cluster and vast rural emptiness looks identical to a state with uniform coverage).
The dot map preserves the original spatial resolution of the data while making density patterns immediately visible through point clustering. It is the correct chart when the story is where, not how many in aggregate — that distinction is the design decision.
A Choropleth Map — the most common alternative — aggregates values by administrative unit (state, county) and encodes them as fill colour. For this data, it would collapse every library in Massachusetts into a single state-level count, hiding the fact that Boston has dozens of branches within a few square miles while rural western Massachusetts has almost none. The choropleth answers "how many per state?" The dot map answers "where exactly?"
A Bubble Map would aggregate counts by city or county centroid, losing the precise location information and misrepresenting the data as a comparison of totals rather than a distribution of individual points. When you have real coordinates for individual events, using them is always more honest than throwing them away to draw circles.
Dot maps do not encode quantity at a point — a dot cluster of twelve libraries and a dot cluster of forty look similar unless you zoom in and count. They are not suitable for retrieving precise counts or comparing exact values between regions. They also suffer from overplotting at dense scales: when many points share near-identical coordinates, they merge visually and the underlying count is lost.
This implementation mitigates overplotting with a small dot radius, semi-transparency on the SVG layer, and zoom — the viewer can drill into any dense cluster to separate individual points. But the limitation is structural: if the message requires exact counts, the dot map hands off to a bar chart or table.
FT Visual Vocabulary — Spatial (Distribution) category. "Use a dot map when you want to show where individual things are located and how they cluster geographically." Abela quadrant: Distribution (geographic). Tufte principle applied: the base map provides essential geographic context — it is not decoration; removing state boundaries would make clustering uninterpretable. But the choropleth fill is deliberately absent — every pixel of colour on the base layer encodes only topology, not data. All data encoding is in the dot positions. The one design decision worth knowing: dot radius is fixed at 3px regardless of zoom level, so density perception at any zoom scale reflects true geographic concentration rather than artificially enlarged points.