Breaking Anonymity at Scale: Re-identifying the Trajectories of 100K Real Users in Japan

2025-06-05

AI Summary

This study identifies critical privacy vulnerabilities in the Japanese Yjmob100k anonymized trajectory dataset (100,000 users), demonstrating that conventional anonymization techniques fail against a synergistic re-identification attack leveraging population density patterns, spatiotemporal structural correlations, and individual activity profiles. The authors present the first systematic evidence that large-scale, end-to-end identity and spatiotemporal trajectory re-identification is feasible using only publicly available geographic information and intrinsic trajectory statistics—without auxiliary external data. Their method integrates geospatial analysis, temporal behavioral modeling, and multi-source structural inference, challenging foundational assumptions of trajectory anonymization security. Experiments successfully recover numerous users’ true identities and fine-grained spatiotemporal trajectories, empirically revealing the dataset’s insufficient anonymization strength. These findings provide crucial theoretical insights and empirical validation for designing privacy-enhancing trajectory publishing mechanisms.

Technology Category

• Machine Learning: Privacy
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Mobility traces represent a critical class of personal data, often subjected to privacy-preserving transformations before public release. In this study, we analyze the anonymized Yjmob100k dataset, which captures the trajectories of 100,000 users in Japan, and demonstrate how existing anonymization techniques fail to protect their sensitive attributes. We leverage population density patterns, structural correlations, and temporal activity profiles to re-identify the dataset's real-world location and timing. Our results reveal that the anonymization process carried out for Yjmob100k is inefficient and preserves enough spatial and temporal structure to enable re-identification. This work underscores the limitations of current trajectory anonymization methods and calls for more robust privacy mechanisms in the publication of mobility data.

Problem

• Re-identifying anonymized user trajectories in mobility data
• Exposing flaws in current trajectory anonymization techniques
• Highlighting privacy risks in large-scale mobility datasets

Innovation

• Leveraging population density patterns for re-identification
• Using structural correlations to analyze trajectories
• Applying temporal activity profiles to reveal identities