📝 SummarXiv

Abstract

Spatial data are gaining increasing importance in many areas of research. Particularly spatial health data are becoming increasingly important for medical research, for example, to better understand relationships between environmental factors and disease patterns. However, their use is often restricted by legal data protection regulations, since georeferenced personal information carries a high risk of re-identification of individuals. To address this issue, what are called geomasking methods are applied to guarantee data protection through targeted displacement of individual data points, while simultaneously maintaining analytical validity within a tolerable range. In the current literature the degree of anonymity of such anonymized georeferenced datasets is often measured by the so-called metric of spatial k-anonymity. However, this metric has considerable shortcomings, particularly regarding its resilience against realistic data attack scenarios. This article classifies the potential data attack scenarios in the context of anonymized georeferenced microdata and introduces appropriate metrics that enable a comprehensive assessment of anonymity adapted to potential data attack scenarios.