PRIVVIZ AND DPVIZETL: ARCHITECTING DIFFERENTIAL PRIVACY IN DATA VISUALIZATION PIPELINES

Abstract

In the age of data democratization and shared analytics, visualizing sensitive data without violating individual privacy has become a pressing challenge. This article introduces two novel frameworks: PrivViz (Privacy-Aware Visualization Framework), a high-level architectural blueprint for enabling differential privacy in visualization environments, and DPVizETL (Differential Privacy Visualization ETL Framework), a practical pipeline model designed to operationalize privacy-preserving data transformation for visual analytics. From theory to implementation, this paper explores their design principles, data engineering considerations, and real-world applicability, offering Data Engineers a structured path toward responsible visual analytics. The frameworks proposed aim to fill the gap between privacy-preserving data publishing and usable, interpretable visual outputs in high-stakes domains such as healthcare, finance, and public policy.