Speaker
Description
In order to fulfill the interoperability requirement for FAIR research data, (meta)data need to comply with a community-agreed-upon language.
In the NOMAD Archive, materials science data are collected from heterogeneous sources, spanning synthesis, experimental characterization, and computations for modelling and analysis. This diversity necessitates flexible storage options, allowing users to create customized metadata schemas.
We present a strategic plan and a first version of a taxonomy for materials properties. A taxonomy is a terminology with structure, i.e., concepts are organized hierarchically in classes and subclasses. Crucially, it is expressed following the OWL2 (Web Ontology Language, version 2) which helps establish a standardized framework. This framework ensures that data from different sources are interoperable, meaning they can be combined, compared, and processed consistently across different systems and platforms.
The NOMAD taxonomy includes for each term: i) a curated, human-understandable definition, ii) a controlled list of synonyms, iii) the specification of the physical dimensions, and iv) the expected shape (tensorial rank). This taxonomy not only shapes the graphical interface of the NOMAD browser but also offers a controlled pool of predefined concepts. These can be reused in custom schemas, providing semantic guidance for newly defined metadata. By adhering to OWL2 standards and exportable in formats like RDF and TTL, our taxonomy facilitates seamless integration with linked data systems, enhancing global data exchange and collaboration within the materials-science community.
Finally, the taxonomy, so far limited to physical properties (i.e., the measured quantities), will be expanded to the description of the materials whose properties are measured and the characterization of experimental or computational methods that are used for performing the measurements.