Structuring superconductor data with ontology: reproducing historical datasets as knowledge bases

Applying historical human-readable databases to recent data-driven science is a natural concept. However, this cannot be realized by simply converting a database into a tabular format because the meaning of each table column and the relationships between columns need to be rewritten in a machine-readable format. In particular, eliminating implicit notations that can only be understood by experts in the fields covered by each database and making them machine-readable under a unified academic system is necessary when integrating data across fields with a view to solving specific social issues and social implementation. In this study, we constructed a superconducting materials ontology for SuperCon, a legacy superconductor database that was recently republished as a datasheet, based on the well-known Basic Formal Ontology (BFO) top ontology, and designed a schema for material composition, structure, properties, and processes, among others. Using this schema, we constructed and published the Resource Description Framework (RDF) for each instance in the SuperCon datasheet. We also discuss the machine-readable format of data common to materials science discovered in this process.