The AR Vis project is a general-purpose interactive data visualization platform for collaborative interaction with scientific data. The platform is designed for augmented reality displays of data supporting multi-user interaction and simulations. Methods developed include a procedural pipeline for data culling, modeling, visualization, and porting to multiuser augmented reality. A prototype interactive visualization application demonstrates the system via visualization and simulation of magnetic fields. The magnetic field visualizations are attached to physical objects or embedded in the environment. The invisible magnetic fields are transformed into tangible models of nano and geospatial scales magnetic phenomena accessible to a user’s full body (embodied) interaction. The project seeks to make a significant contribution to scientific visualization. Extending beyond the cognitive impact of traditional scientific visualization, the goal of the AR Vis platform is to additionally leverage human perception and spatial cognition and make data patterns tangible, manipulable and more accessible. In supporting augmented information cognition in scientists and learners, AR Vis design supports data discovery and learning. The prototype implementation uses physics data modeling of the invisible and largely intangible forces of magnetism across different scales. The project yields both a prototype platform and develops a data visualization pipeline. Both demonstrate a substantial and concrete implementation and demonstration of AR Vis techniques.