Privacy guarantees for personal mobility data in humanitarian response

Area of Science:

• Computer Science
• Data Privacy
• Humanitarian Informatics

Background:

• Personal mobility data is crucial for humanitarian response but raises privacy concerns.
• Aggregated data can still expose sensitive individual movement patterns.

Purpose of the Study:

• To develop and validate a privacy-preserving algorithm for releasing mobility data.
• To enable the responsible use of mobility data in high-stakes policy decisions.

Main Methods:

• Developed a differentially private algorithm for mobility matrix construction.
• Derived formal privacy and accuracy bounds for the algorithm.
• Applied the algorithm to real-world mobile phone data from Afghanistan and Rwanda.

Main Results:

• Demonstrated the algorithm's effectiveness in enabling privacy-preserving mobility data analysis.
• Showcased applications in pandemic response and humanitarian aid distribution.
• Identified key trade-offs between privacy and accuracy in practical implementation.

Conclusions:

• The proposed approach facilitates the responsible use of personal mobility data in humanitarian contexts.
• Formal privacy guarantees are essential for protecting individuals while leveraging mobility insights.
• Balancing privacy and accuracy is critical for effective policy-making with sensitive data.