K-Area: An Efficient Approach to Approximate the Spatial Boundaries of Mobility Data with k-Anonymity

Mobility datasets, being by nature potent in utility and complexity, are hard to work with when privacy has to be preserved. Existing solutions to balance utility and privacy are very specific to certain use case or dataset types, and usually strive to provide an absolute privacy while disregarding computational efficiency. K-area is an efficient method that uses geometric operations to calculate the boundaries of movement profiles that guarantee a certain degree of anonymity and exclude areas where privacy is at risk. It is applicable to most types of mobility datasets, as it only requires a set of Global Positioning System (GPS) points tagged to an identifier. K-area provides the largest areas of the dataset, which all validate a geometric k-anonymity condition. By already providing a level of indistinguishability, these areas are the perfect starting point for many applications.