Visualizing the Risks of De-anonymization in High-Dimensional Data

De-anonymization is a type of privacy breach in which an adversary identifies an individual in an anonymized dataset, thus gaining access to the individual’s private and sensitive data. To that end, data controllers who own such data and wish to distribute it to third parties have to be aware of the risks associated with their dataset and implement the appropriate anonymization measures. In order to conduct de-anonymization risk analyses, data controllers are in need of methods that can highlight risks, aid in regulatory compliance, and guide them through the anonymization process. The main drawbacks of existing de-anonymization analysis solutions are their insufficiency to deal with high-dimensional data as well as their limited visualization possibilities for users to evaluate the results of the analyses. The presented work introduces four novel risk analysis methods for four types of complex, high-dimensional data through intuitive and interactive visualizations while also improving on existing ones.