If every place belonged to whichever facility was nearest, what would the map look like?
Place several clinics, schools or wells on an empty plane. Around each point lies a region containing every location closer to that point than to any other. The borders occur where distance to two neighbouring points is equal.
These are Voronoi regions, also called Thiessen polygons in many geographic applications.
Voronoi geometry answers one beautifully simple question: which point is nearest?
Quick Read: The Mechanism
GENERATOR POINTS → NEAREST-DISTANCE RULE → EQUIDISTANT BOUNDARIES → NON-OVERLAPPING TERRITORIES
1. Every Location Gets One Nearest Generator
Ignoring exact ties on boundaries, each location is assigned to the nearest point. The resulting polygons cover the plane without gaps or overlaps.
2. Boundaries Are Equidistant
A border between two cells lies where locations are equally distant from the two corresponding generator points. The geometry is determined entirely by the point locations.
3. Voronoi Is a Model of Proximity, Not Actual Service
The nearest clinic may be full, expensive or separated by a river. The nearest school may not be the assigned school. Voronoi regions model geometric proximity, not capacity, eligibility or behaviour.
4. Euclidean Geometry Is the Default
Classic Voronoi cells use straight-line distance. In cities, Network Distance may produce very different practical catchments.
5. Weighted Voronoi Changes the Competition
If facilities differ in size, quality or attractiveness, weighted variants can give stronger generators larger effective territories. This moves the model closer to real competitive geography while adding assumptions.
6. Voronoi Is Not Central Place Theory
Central Place Theory owns thresholds, ranges and service hierarchies. Voronoi geometry owns nearest-point partitioning.
7. Voronoi Is Not Location–Allocation
Location–Allocation chooses facility locations under objectives and constraints. Voronoi regions take the points as given and partition surrounding space by proximity.
8. Thiessen Polygons in Environmental Geography
Rain gauges can be assigned polygons so each part of a catchment takes the rainfall measured at its nearest gauge. This is simple and transparent, but it creates abrupt boundaries in a phenomenon that may vary continuously.
9. Delaunay Triangulation Is the Dual
Connect generator points whose Voronoi cells share a boundary and a Delaunay triangulation emerges. This dual structure is useful for neighbourhood and surface algorithms.
10. Primary Geography: Who Is Closest?
Put three water stations on a playground map and ask pupils to colour every place according to its nearest station. The resulting boundaries reveal Voronoi logic intuitively.
11. Secondary Geography: Compare Geometric and Real Catchments
Students can draw nearest-school Voronoi cells and compare them with actual travel routes or administrative catchments. The differences reveal barriers, policy and network effects.
12. Advanced Geography: Metric Choice Changes the Tessellation
Changing from Euclidean distance to travel time, weighted distance or network cost changes the territory. A Voronoi diagram is therefore always tied to a distance definition.
13. Singapore Example: MRT Stations
Euclidean Voronoi cells around MRT stations provide a first approximation of nearest-station territories. Actual station choice can differ because entrances, walking networks, bus feeders, line destinations and interchanges alter practical access.
14. Healthcare Example
Nearest-clinic polygons can reveal simple geographic coverage gaps, but they ignore capacity, specialisation, opening hours and patient preference.
15. Ecology Example
Voronoi cells can approximate territories around nests or plants, but biological territory depends on habitat, competition and behaviour rather than pure geometry.
16. Hostile Test: “Everyone Inside This Polygon Uses This Facility”
No. The polygon says only that the facility is nearest under the chosen metric. Behaviour requires additional evidence.
17. Where Voronoi Reasoning Breaks
- Nearest-equals-chosen: confusing geometric proximity with behaviour.
- Capacity blindness: assuming all generators can serve unlimited demand.
- Barrier blindness: ignoring rivers, roads and restricted access.
- Metric blindness: failing to state what distance defines nearest.
- Abrupt-surface fiction: assigning continuous environmental values as sharp polygon steps.
- Policy blindness: confusing nearest territory with official catchment.
18. Ten Questions for Voronoi Regions
- What are the generator points?
- What distance metric is used?
- Are all generators equally attractive?
- Do barriers matter?
- Does network travel differ from straight-line distance?
- Does capacity matter?
- Are official territories different?
- Is the phenomenon continuous or discrete?
- Would weighting improve the model?
- What does the polygon legitimately claim?
19. Where This Fits
Location–Allocation owns optimal facility placement. Central Place Theory owns service hierarchy and range. This article owns nearest-generator partitioning of geographic space.
The Idea to Keep
Voronoi regions are powerful because they make one assumption perfectly clear: nearest wins.