Leveraging the Schema.org Vocabulary to Create an Actionable Metadata Representation for Geospatial Data and Computing Resources

This resource contains slides for the AGU Fall Meeting 2023 presentation (#IN23A-07) in San Francisco on Dec 12.
Session: IN23A: Advancing Open Science: Emerging Techniques in Knowledge Management and Discovery II Oral

Effective response to global crises relies on universal access to scientific data and models, understanding their attributes, and representing their interconnectivity to facilitate collaborative research and decision making. In the age of distributed data, geospatial researchers frequently invest significant time searching for, accessing, and working to understand scientific data. This often leads to the recreation of existing datasets as well as challenges in determining methods for accessing, using, and ultimately establishing connections between resources. In recent years, following FAIR and CARE principles, there is an emerging practice to leverage structured and robust metadata to accelerate the discovery of web-based scientific resources and products. This practice assists users in not only discovery, but also in understanding the context, quality, and provenance of data, as well as the rights and responsibilities of data owners and consumers. It also empowers organizations to leverage their data more effectively and derive meaningful insights from them. Doing so, however, can be difficult, especially when diverse resources needed for scientific applications may be spread across multiple repositories or locations. We present a solution for leveraging the Schema.org vocabulary along with various web encodings such as the Resource Description Framework (RDF) with JSON-LD to create an actionable, curated catalog of scientific resources ranging from spatio-temporal data to software source code. We explore how resources of various types and common scientific formats, such as multidimensional, software containers, source code, and spatial features, which are stored across various repositories and distributed cloud storage, can be described and cataloged. Recognizing the impracticality of manually cataloging metadata, we have developed generic capabilities to automatically extract metadata for such resources, while empowering scientists to provide additional context. By incorporating comprehensive metadata, the exploration of diverse data relationships can be realized to gain insight into gaps and opportunities to improve the connectivity between science communities.