Privacy-preserving data warehousing

Abstract

Data warehouses are an important element of business intelligence and decision support in many companies and inter-organisational data infrastructures. However, when personal information of individuals is concerned, it is critical to provide sufficient protection mechanisms in order to preserve privacy. In addition to classical access control, database anonymisation is an important element of an encompassing strategy for privacy-preserving data storage. This article gives an overview on selected anonymisation concepts and techniques and investigates if they are suitable for a data warehouse context. Furthermore, a process of privacy-preserving data integration and provisioning is presented and the impact of architecture, privacy criteria, and further parameter choices is discussed. Finally, we experimentally compare the impact of these parameters on data utility after anonymisation in several experiments on multiple datasets and derive corresponding recommendations.