The lack of a standard for the exchange data for scientific visualization is a big hurdle in the interoperability among applications. The requirement to develop support for many diverse file formats demands human resources that could otherwise be spent more efficiently on real research topics. The vision of a unified data model is to provide a common denominator for at least a wide range of data types. This article presents the treatment of frequently occurring data types within the context of a data model based on the mathematics of fiber bundles, casting data in a six-level hierarchy. This scheme allows the derivation of a taxonomy of data and operations on them, as exemplified in this presentation.