3D skeleton based gait representations have been widely applied to various areas with many advantages, while most existing models lack the ability to provide intuitive or human-understandable explanations for the effectiveness of learned representations. In this work, we propose a general 3D skeleton based Gait Visualization and Analysis (3DGVA) platform, which can not only interactively present unified skeleton-based gait patterns but also provide the visualization of feature-based gait ranking for qualitative evaluation. Moreover, 3DGVA synergizes large language models (LLMs) to perform automatic gait analysis to facilitate users’ model assessment with explainable reasoning. Our 3DGVA platform is able to visualize and analyze different state-of-the-art skeleton-based models for person re-identification, gait score prediction, disease prediction tasks, and can be potentially extended to more related interdisciplinary applications.