A differential privacy-based privacy-preserving data publishing algorithm for transit smart card data

This manuscript is focused on transit smart card data and finds that the release of such trajectory information after simple anonymization creates high concern about breaching privacy. Trajectory data is large-scale, high-dimensional, and sparse in nature and, thus, requires an efficient privacy-preserving data publishing (PPDP) algorithm with high data utility. This paper describes the investigation of the publication of non-interactive sanitized trajectory data under a Differential Privacy (DP) definition. To this end, a new prefix tree structure, an incremental privacy budget allocation model, and a spatial-temporal dimensionality reduction model are proposed to enhance the sanitized data utility as well as to improve runtime efficiency. The developed algorithm is implemented and applied to real-life metro smart card data from Shenzhen, China, which includes a total of 2.8 million individual travelers and over 220 million records. Numerical analysis finds that, compared with previous work, the proposed model outputs sanitized dataset with higher utilities, and the algorithm is more efficient and scalable.