Making Tabular Data FAIR: Applying CSV on the Web (CSVW) for Semantic Interoperability in Computational Science and Engineering

Tabular data are ubiquitous in Computational Science and Engineering, serving as both input and output for complex simulations and data processing. They are also the data behind most scientific visualizations— plots and figures used to communicate and substantiate research findings. A variety of file formats have evolved for tabular data exchange, ranging from simple text-based formats (e.g., space-delimited files, comma-separated values (CSV), etc.) to highly structured binary formats such as HDF5 or NetCDF. While many of these formats support some degree of machine readability—such as data types and table schema information—they often lack a universal, machine-interpretable semantic description of the data content itself, including the physical quantities represented and their corresponding units. To address this gap, we explore the implementation of the CSV on the Web (CSVW) standard, which provides a framework for enriching tabular data with semantic metadata to enhance their Findability, Accessibility, Interoperability, and Reusability (FAIR). In this contribution, we report on our experience applying CSVW to provide machine-readable, semantically rich metadata for recent datasets in the field of Computational Materials Science. Our work highlights the integration of domain ontologies such as QUDT (Quantities, Units, Dimensions and Data Types Ontologies), and discusses practical aspects of tool support in Python and other scripting environments. The presented approach demonstrates how CSVW may effectively bridge the gap between conventional data formats and FAIR-compliant semantic data publication in computational research.