Privacy-Preserving Analytics of Human Mobility Data: Investigating the Gap between State-Of-The-Art Privacy Methods and Real-Life Utility Requirements
Year: 2024
Type: dissertation
Abstract: In recent years, human mobility data has been increasingly collected and stored, mainly due to the ubiquitous use of smartphones, producing a constant stream of cellular network and GPS data. This provides an unprecedented resource for useful applications in various fields, such as urban and traffic planning or disaster response. At the same time, as more and more data becomes available, individuals' privacy is increasingly at risk. As a result, there is a growing body of research on privacy-enhancing technologies. Such methods, however, come with an inherent utility-privacy trade-off, where the usefulness of the data is traded for privacy. State-of-the-art techniques like methods ensuring differential privacy are not (yet) applied to a large extent in practical settings. This thesis examines privacy-preserving technologies for human mobility data in terms of their usefulness for modeling and analysis projects, with a particular focus on real-world practices. A differentially private mobility data report is proposed to serve practitioners as a tool to provide typical mobility data statistics with high privacy standards. Synthetic data generation methods are explored in-depth and can thus far not be recommended for usage in practice. For this purpose, existing methods for evaluating the utility of synthetic or otherwise privacy-enhanced mobility data are critically reviewed with regard to their suitability for various application scenarios and supplemented by more practice-relevant measures. Analytics of mobility data in a privacy-preserving manner remains an ongoing issue requiring context-specific solutions.:Introduction 1.1 Motivation 1.2 Scope 1.3 Contribution and Thesis Outline 2 Background and Related Work 2.1 Human Mobility Data 2.1.1 Spatio-Temporal Trajectories 2.1.2 Data Sources 2.1.3 Applications 2.1.4 Utility Measurement 2.2 Technical Anonymization 2.2.1 Basic Privacy Mechanisms 2.2.2 k-Anonymity 2.2.3 Di!erential Privacy 2.2.4 Synthetic Data 2.2.5 Legal Perspective 3 Collection, Usage and Privacy of Mobility Data in the Enterprise and Public Administrations 3.1 Introduction 3.2 Publication 3.3 Conclusion 4 Towards Mobility Reports with User-Level Privacy 4.1 Introduction 4.2 Publication 4.3 Conclusion 5 Generative Models for Synthetic Urban Mobility Data: A Systematic Literature Review 5.1 Introduction 5.2 Publication 5.3 Conclusion 6 Reconsidering Utility: Unveiling the Limitations of Synthetic Mobility Data Generation Algorithms in Real-Life Scenarios 6.1 Introduction 6.2 Publication 6.3 Conclusion 7 Conclusion and Discussion 7.1 Mobility Data Analytics Applications in Need of Technical Anonymization 7.2 Utility Measures 7.3 Differential Privacy and Mobility Data 7.4 Synthetic Mobility Data 7.5 Beyond Technical Anonymization References
Author: Kapp, Alexandra
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6986581844
