As experimental and simulation capabilities have continually increased, the need has developed to manage the resulting large data sets. Techniques such as serial sectioning lead to data sets on the order of gigabytes. Likewise, large finite element simulations can involve millions of elements and also produce gigabytes of output. Developed for the U.S. Navy, the Materials Atlas is a database and storage system for managing large data sets arising from various sources. Specifically, the project targets titanium and stainless steel data. Data sets are stored using a model based on HDF5, a scientific data format developed at the National Center for Supercomputing Applications. Information about the data, including material chemistry, crystal structure and phase definition, is stored separately in a relational database. Ultimately, data sets can be accessed and viewed according to data source (experiment, simulation, etc.), material or material phases, or other meta-data. The author will discuss the schema design, the software implementation and user experience.