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

Anonymizing microdata requires balancing disclosure risk reduction with the preservation of data utility. Traditional evaluations often rely on single measures or two-dimensional risk-utility (R-U) maps, but real-world assessments involve multiple, often correlated, indicators of both risk and utility—a fundamentally multivariate problem that pairwise comparisons fail to capture both efficiently and completely. We compare six visualization approaches for the simultaneous evaluation of multiple risk and utility measures: heatmaps, dot plots, composite scatterplots, parallel coordinate plots, radial profile charts, and principal component analysis (PCA)-based biplots. We introduce blockwise PCA for composite scatterplots and joint PCA for biplots that simultaneously reveal method performance and measure interrelationships, and apply systematic Pareto-optimal method identification across all approaches where applicable, with dominance assessed in the original composite score space. Our comparison shows that no single approach dominates across all criteria: PCA biplots perform well on the analytical criteria, in particular for revealing how the risk and utility measures relate to one another, while composite scatterplots do so on scalability and applicability and are the only approach that can display the Pareto front. Combining complementary visualizations provides the most complete basis for evaluating the risk-utility trade-off.