DSD: The Data Source Description Vocabulary

Training machine learning models, especially in producing enterprises with numerous information systems having different data structures, requires efficient data access. Hence, standardized descriptions of data sources and their data structures are a fundamental requirement. We therefore introduce version 4.0 of the Data Source Description Vocabulary (DSD), which represents a data source in a standardized form using an ontology. We present several real-world applications where the DSD vocabulary has been applied in recent years to demonstrate its relevance. An evaluation against the FAIR principles highlights the scientific quality and potential for reuse of the DSD vocabulary.