As large data sets involving sensitive personal information such as health care data become more desired by researchers, the need to develop methods of protecting the identities of individuals becomes increasingly important. In this paper, geographic methods of anonymizing large sensitive data sets are discussed. Given an input dataset that includes geographic information, noise can be applied to the dataset to enhance anonymity. These methods are known as geographic perturbation. Geographic perturbation algorithms that use nearest neighbour information are investigated, with their resulting data utility and level of anonymization compared. The three methods investigated are Voronoi snapping, Voronoi snapping with noise and random Voronoi snapping.