ODCalibrator: An Interactive Visualization System for OD Traffic Flow Calibration in Microscopic Traffic Simulations

Traffic simulation is essential for evaluating transportation policies, infrastructure changes, and mixed traffic scenarios with human-driven and autonomous vehicles, yet its reliability critically depends on accurate calibration of origin–destination (OD) traffic demand. Traditional calibration workflows rely on trial-and-error adjustments and static numerical outputs, making it difficult for analysts to interpret error patterns, understand OD-to-link relationships, and efficiently refine traffic demand in complex urban networks. To address these challenges, we conducted a design study with transportation simulation experts to characterize the OD calibration workflow and derive key analytical tasks, which informed the development of ODCalibrator, an interactive visualization system that supports human-in-the-loop calibration through coordinated views, visual diagnostics, and iterative adjustment capabilities. We demonstrate its utility via a narrative usage scenario and domain expert feedback, showing that the system enables analysts to quickly identify error-prone regions, explore the effects of OD adjustments, and leverage domain expertise to produce more efficient and interpretable calibration outcomes for urban traffic simulations.