Data discovery with DATS: exemplar adoptions and lessons learned. (8th December 2017)

• Record Type: Journal Article
• Authors: Gonzalez-Beltran, Alejandra N; Campbell, John; Dunn, Patrick; Guijarro, Diana; Ionescu, Sanda; Kim, Hyeoneui; Lyle, Jared; Wiser, Jeffrey; Sansone, Susanna-Assunta; Rocca-Serra, Philippe
• Is Part Of: Journal of the American Medical Informatics Association. Volume 25:Number 1(2018)
• Page Start: 13
• Page End: 16
• Publication Date: 2017-12-08
• Subjects: data discovery -- data model -- search engine -- metadata; Medical informatics -- Periodicals; Information Services -- Periodicals; Medical Informatics -- Periodicals; Médecine -- Informatique -- Périodiques; Informatica; Geneeskunde; Informatique médicale; Computer network resources; Electronic journals; 610.285
• DOI: 10.1093/jamia/ocx119
• Languages: English
• ISSNs: 1067-5027

Abstract

Abstract: The DAta Tag Suite (DATS) is a model supporting dataset description, indexing, and discovery. It is available as an annotated serialization with schema.org, a vocabulary used by major search engines, thus making the datasets discoverable on the web. DATS underlies DataMed, the National Institutes of Health Big Data to Knowledge Data Discovery Index prototype, which aims to provide a "PubMed for datasets." The experience gained while indexing a heterogeneous range of >60 repositories in DataMed helped in evaluating DATS's entities, attributes, and scope. In this work, 3 additional exemplary and diverse data sources were mapped to DATS by their representatives or experts, offering a deep scan of DATS fitness against a new set of existing data. The procedure, including feedback from users and implementers, resulted in DATS implementation guidelines and best practices, and identification of a path for evolving and optimizing the model. Finally, the work exposed additional needs when defining datasets for indexing, especially in the context of clinical and observational information.