Voronoi Diagrams

Voronoi Diagrams: Partitioning the World Based on Proximity

Imagine a city with several fire stations. Every resident faces a question: 'If a fire breaks out, which fire station should I turn to?' The simplest answer is: 'Go to the one closest to you.'

Consequently, we can divide the entire city into distinct zones: people in one zone are closest to fire station A; those in another are closest to fire station B; and those in yet another are closest to fire station C. Ultimately, every location is assigned to the fire station nearest to it.

If you map out these zones, you get a striking geometric pattern known as a Voronoi diagram.

It All Starts with a Few Points

The foundation of a Voronoi diagram is quite simple: a set of points scattered across a plane. These points can represent anything—fire stations, hospitals, shops, schools, cities, or even trees. We refer to these points as generating points or sites.

Suppose we have three points: A, B, and C. For every location on the plane, we identify the point closest to it. For instance, all locations within a certain area surrounding point A are closer to A than to any other point. Similarly, points B and C each claim their own respective zones. The entire plane is thus partitioned into three parts: the zone for A, the zone for B, and the zone for C. Each of these zones is called a Voronoi cell or Voronoi region. Each generating point corresponds to one Voronoi region; 10 points will typically yield 10 regions, and 100 points will result in 100 regions.

So, where do the boundaries lie?

This is the most crucial aspect of a Voronoi diagram.

Suppose you have only two points, A and B. The question then becomes: when is a location equidistant from both points? The answer: when it lies exactly halfway between them. Consequently, the straight line running right between points A and B serves as the dividing line. Every point on this line is at an identical distance from both A and B, while the space on either side of the line belongs to either A or B. In mathematical terms, the set of all points equidistant from two given points forms a perpendicular bisector. Thus, the boundaries in a Voronoi diagram are not drawn arbitrarily; they are derived from a very simple mathematical rule: identify all locations that are equidistant from two given seed points.

What happens when you have three points? If we add a third point, C, then A, B, and C all begin to compete for their own regions. For any given location, the question asked is: 'Which point am I closest to?' Ultimately, the entire plane is partitioned into three distinct regions, with the boundaries between them consisting of lines that are equidistant from the respective seed points.

As the number of points increases, the pattern formed by these boundaries becomes increasingly complex. Whether you start with 3 points and scale up to 10, 50, or even 100, you will observe the entire plane rapidly transform into an intricate network of polygons. Yet, upon closer inspection, you will see that each region remains centered around a single seed point—the underlying rule has never changed.

Why does it look so much like cells?

If you randomly generate a large number of points, a Voronoi diagram often produces a very familiar visual pattern: cells. It creates a collection of irregular polygons packed tightly together, where each region has its own central point, and the boundaries connect seamlessly with the surrounding areas.

This is why Voronoi diagrams often evoke images of biological cells, honeycombs, cracked earth, the texture of animal skin, foam, or even city blocks.

However, there is an important distinction to be made: just because things look similar does not mean they are all generated by Voronoi diagrams. A Voronoi diagram is, first and foremost, a mathematical rule for partitioning space. While some structures in nature may relate to this geometric principle, others arise from entirely different physical processes.

Why are the boundaries always so neat? This is the beauty of a Voronoi diagram. Consider two points, A and B; the boundary between them consists of all locations equidistant from both—making the boundary, by definition, the perpendicular bisector of the line segment AB. When more points are added, these perpendicular bisectors intersect, ultimately forming a network of polygons defined by straight lines. Thus, while a Voronoi diagram may appear random, it is actually governed by a very precise geometric structure.

It Can Represent More Than Just Fire Stations

Let’s return to our initial example: fire stations. Suppose a city has several fire stations—A, B, C, D, and so on. If we represent each fire station as a point and generate a Voronoi diagram, each resulting Voronoi region can be understood as the area for which a specific fire station is primarily “responsible.” If a resident lives within the region associated with station A, then A is the closest fire station to them; if they live in the region for station B, then B is the nearest one. Thus, what appears to be an elegant geometric diagram is, in fact, a direct representation of a very real-world problem.

Voronoi diagrams aren’t limited to fire stations; the same principle applies to many other scenarios. Which hospital is closest to me? Which store serves this particular area? Which school is nearest to a student? Which distribution center is responsible for a given zone? Which mobile base station is closest to a specific phone? Far from being merely a beautiful mathematical pattern, the Voronoi diagram is a highly practical tool for spatial partitioning.

At its core, the Voronoi diagram is all about one concept: proximity (or “nearest”). Given a set of points, for any location on a plane, one might ask: “Which point is closest to me?” Answering this question determines which region that location belongs to. By mapping the answers for every possible location, a Voronoi diagram emerges.

You can even think of a Voronoi diagram as a mathematical machine. It takes a set of points as input and, for every location on the screen, asks: “Which point are you closest to?” If the answer is A, that location is assigned to A; if it’s B, it’s assigned to B; if it’s C, it’s assigned to C, and so on. Ultimately, the entire space is automatically partitioned.

Order Can Emerge from Random Points

If we scatter a number of points randomly—with their positions entirely arbitrary—and then apply the simple rule that “every location belongs to the nearest point,” a highly organized geometric structure emerges. In other words, random inputs can give rise to an ordered partitioning of space.

This is distinct from a fractal; a Voronoi diagram is not a fractal. It lacks infinite recursion, infinite self-similarity, or structures that repeat as you zoom in. Its core essence is the partitioning of space.

Each Point Acts as a “Center”

You can visualize a Voronoi diagram as a collection of points competing for space. Each point asserts, “This area is closest to me,” while another counters, “No, it’s closer to me.” Boundaries eventually form between them, and the entire plane is carved into countless distinct regions based on this competition for “nearest proximity.”

What’s even more fascinating is that changing the rules transforms the world itself. A standard Voronoi diagram relies on Euclidean distance—the familiar, straight-line distance. For instance, if one point is 3 meters away from you and another is 5 meters away, the first point is clearly closer.

However, the real world is rarely that simple. Consider a city: the hospital that is closest in terms of straight-line distance might not be the one you can reach fastest by car, as factors like roads, rivers, mountains, and traffic come into play. Thus, “distance” is not limited to the Euclidean definition. If you alter the definition of “distance,” the resulting Voronoi diagram changes accordingly. This is the beauty of mathematics: changing a single rule can reshape an entire world.

A Simple Question That Creates an Entire Map

When you look at a beautiful Voronoi diagram, don’t just see it as a collection of randomly generated, colorful polygons. What you are actually seeing is the answer to the question, 'Which point is closest?' for every location in the entire space. Each region answers the question, 'To whom does this area belong?' Each boundary marks the point where a different site becomes the nearest one. And every vertex represents a unique geometric relationship between multiple regions.

The story of a Voronoi diagram can be summed up in a single sentence: Give me a set of points, and I can partition the entire world based on which point is closest to any given location. There is no complex recursion, no mysterious formula, and no artificial design—just points, distance, and proximity. Yet, when these simple rules are applied across an entire plane, a complex and beautiful geometric world emerges.

Exploring the World of Mathematics Through 'Proximity'

Voronoi diagrams illustrate how distance shapes space.

You might also explore the Sierpinski triangle, where a simple triangle is transformed through repeated 'punching out' of holes, creating an infinitely recursive pattern.

Or, take a look at the Koch snowflake, which shows how a simple straight line can evolve into an infinitely complex snowflake through continuous modification of its edges.

You could also investigate the Hilbert curve—a curve that, through endless recursion, gradually fills an entire two-dimensional space.

Or perhaps the dragon curve, which demonstrates how a simple line can be folded repeatedly to form a mysterious, dragon-like structure.

Voronoi diagrams, by contrast, work quite differently. They don't rely on infinite recursion; instead, they simply ask: 'What is closest to this point?' This single, straightforward question is enough to transform an entire space into a stunning mathematical map.

Feel free to discover more fascinating mathematical patterns, or check out the Wikipedia entry on Voronoi diagrams for further information.