A Utility-Aware Visual Approach for Anonymizing Multi-Attribute Tabular Data.

Area of Science:

• Data privacy and visualization
• Information security and human-computer interaction

Background:

• Public data sharing necessitates robust sanitization to prevent sensitive information disclosure.
• Existing privacy-preserving visualization methods often lack adequate user feedback on data utility.
• Quantifying utility loss during data anonymization remains a significant challenge.

Purpose of the Study:

• To develop an interactive visual interface and data manipulation pipeline for privacy-preserving data handling.
• To enable users to effectively gauge and manage utility loss during data sanitization.
• To integrate and compare established privacy models like syntactic anonymity and differential privacy.

Main Methods:

• Design of a visual interface coupled with a data manipulation pipeline.
• Interactive and iterative data processing for privacy management.
• Integration and comparative analysis of syntactic anonymity and differential privacy models.

Main Results:

• Demonstrated effectiveness of the approach through case studies on diverse datasets.
• Enabled users to visualize and assess the impact of privacy measures on data utility.
• Facilitated a user-centric approach to balancing data privacy and utility.

Conclusions:

• The developed visual interface and pipeline offer an effective solution for privacy-preserving data sharing.
• The approach provides crucial feedback on utility loss, enhancing user trust and data usability.
• Comparative analysis of privacy models aids in selecting appropriate methods for specific use cases.