Given a discrete subset of \(D \subset \mathbb{R}^d\), you can imagine that \(D\) is a realization of a Poisson point process, we associate to each \(x \in D\), its Voronoi cell \(V(x) \subseteq \mathbb{R}^d\), which consists of all the points \(y \in \mathbb{R}^d\), such that \(\| x- y\| < \|z - y\|\) for all \(z \in D \setminus \{x\}\). You can imagine that each point (cellphone location) \(y \in \mathbb{R}^d\) would like to report to the nearest (cellphone tower) point in \(D\), and \(V(x)\) is all the points that report to \(x\). For more information see wiki link.
Here we consider a simple voting model studied in by Quas. Suppose we have \(n\) candidates, and their political positions are represented on a point on the unit interval \([0,1]\), and subsets of \([0,1]\) represents a share of the voters. The support that each candidate has, is given by their Voronoi cell. The candidates go through, \(n-1\) rounds of election, where in each round, the candidate with the least support drops out.