Beyond the Trade-off Curve: Multivariate and Advanced Risk-Utility Maps for Evaluating Anonymized and Synthetic Data

Anonymised and synthetic data are usually judged on a single risk-versus-utility trade-off curve, which collapses a complex, multidimensional question into one line. This preprint proposes richer, multivariate risk-utility maps that show how different risk and utility measures move together, revealing trade-offs the classical curve hides. The maps give data holders a more honest, fine-grained picture for choosing among candidate protected datasets. Developed within the SwissAnon group, the work advances how the quality of anonymised and synthetic data is evaluated and communicated.