At DBA, we are conducting two initiatives: ``visualization analysis of crowd behavior data'' and ``development of multi-agent-based simulation models.'' In the first effort, we used an efficient method based on computer vision algorithms to extract human movements from surveillance camera video data and visualize and analyze crowd behavior. In the second effort, we developed a multi-agent-based simulation model to perform more realistic simulations and parameter modifications using VR technology. DcBA aims to obtain a more accurate model through comparison with ground truth data while complementing the two approaches. The ultimate goal is to apply the developed model to the architectural design process and provide insights into pedestrian flow and behavior.