Panel: Metadata and Knowledge Graphs

CONTENTdm is a digital asset management service from OCLC. Working closely with product and technology colleagues, OCLC Research partnered with five CONTENTdm institutions to investigate how to convert current CONTENTdm metadata into linked structured data, manage that data, create new data, and leverage the structed data to improve the end-user experience of CONTENTdm. Wikibase, the platform used to support WikiData, was selected for this project because it provides a large number of services and features that are critical to this work. It also has a large and engaged community of support. This talk with discuss the framework for the project, provide a technical overview of the data and services, and share participant feedback from the project.

Research Software is key to understand, reproduce and reuse existing scientific results in many disciplines, ranging from Geosciences to Astronomy or Artificial Intelligence. However, research software is usually difficult to find, reuse, compare and understand due to its disconnected documentation (dispersed in manuals, readme files, web sites, and code comments) and a lack of structured metadata to describe it. In addition, research software is no longer isolated code, but a connected network of scripts, libraries, notebooks and containers that represent different aspects and dimensions of a research software project. These problems affect not only researchers, but also students who aim to compare published findings and policy makers seeking clarity on a scientific result. In this talk I will present an overview of the main challenges when creating Knowledge Graphs of Research Software metadata, together with our latest efforts to capture this knowledge from software documentation and code to make it useful to researchers.

Elsevier Health provides a suite of offerings to clinicians and patients both at the point of care and throughout their professional education and training. Our products are backed by a multitude of datasets managed by different teams across the company. The Healthcare Knowledge Graph (H-Graph) was launched in 2017 in an effort to facilitate linking data in the health domain across products and datasets, and remove barriers to cross-product innovation. H-Graph contains high-quality medical knowledge sourced through a combinatorial approach of expert-curated and machine learning extracted relations between medical concepts and their links to source material and associated metadata. In this talk, we will review the progress of H-Graph over the past 5 years, the challenges encountered, and our proposed path forward.