GeoAvatar: A big mobile phone positioning data-driven method for individualized pseudo personal mobility data generation

Abstract

The importance of personal mobility data is widely recognized in various fields. However, the utilization of real personal mobility data raises privacy concerns. Therefore, it is crucial to generate pseudo personal mobility data that accurately reflects real-world mobility patterns while safeguarding user privacy. Nevertheless, existing methods for generating pseudo mobility data, mostly focusing on trip or trajectory generation, have limitations in capturing sufficient individual heterogeneity. To address these gaps, taking pseudo-person(avatar) as ground-zero, a novel individual-based human mobility generator named GeoAvatar has been proposed – which considering individual heterogeneity in spatial and temporal decision-making, incorporates demographic characteristics. Our method utilizes a deep generative model to generate heterogeneous individual life patterns, a variation inference model for inferring individual demographic characteristics, and a Bayesian-based approach for generating spatial choices considering individual demographic characteristics. Through our method, we have achieved generating realistic pseudo personal human mobility data - we evaluated the proposed method based on physical features – obeying common law of human mobility, activity features – showing diverse and realistic activities, and spatial-temporal characteristics – presenting high-accuracy in terms of temporal grid population and od-count, demonstrating its good performance, with both a big mobile phone GPS trajectory dataset from Tokyo Metropolis and a big mobile phone CDR dataset from Shanghai. Furthermore, this method maintains extensibility for broader applications, making it a promising framework for generating pseudo personal human mobility data.