Quantifying Metadata's Value in a Linked Data Landscape: Concepts, Questions, and Methods using SQLite and Python
As discovery products increasingly use linked data ontologies, libraries reflect on traditional metadata's value. Several studies at WVU Libraries explore whether metadata facilitates more precise retrieval than linked data alone. The first uses SQLite in Python to compare precision and relevance of retrieval using subject metadata and equivalent terms from a linked data ontology, EBSCO Discovery Service's "Concept Map." The methodology can be adapted to measure the performance of discovery products. A second study plans to use the newly released WC Entities product. SQLite-based processes are used to quantify the increase in precision retrieval achieved by using attributes in a WC Entity record over the WC Entity label alone. These studies, summarized and presented with examples of code, demonstrate how MARC and other traditional metadata is still valuable to evaluate and enrich linked data for discovery and reinforce the importance of dedicating library resources to its creation and management.
Quantifying Metadata's Value in a Linked Data Landscape: Concepts, Questions, and Methods using SQLite and Python Date: 25 April 2024 Time: 11:00 am – 12:00 pm EDT (UTC -4) Register now
Featuring: Emily Fidelman, Head of Metadata Services, West Virginia University
