Metadata standard for data
More and more libraries purchase datasets from vendors and provide data management services for their users. While these changes bring excitement and opportunities to libraries as a way to expand their services to researchers, data services also bring challenges, notably how to describe and provide access to the data so these data and datasets are discoverable regardless of platforms and reusable by users. Because data have different characteristics than traditional library resources, they require new metadata standards designed specifically for describing and managing data. Those standards can be designed based on the system where data are stored, managed, and discovered, or they can be designed based on the user community that creates and uses data. The seven articles in this special issue of the Journal of Library Metadata assess common practices in creating metadata for data, discuss challenges encountered when creating and sharing metadata developed for different data types, and share the best practices for describing the data identified in the libraries' collections.
How this special issue is organized
First, Manifestations of Metadata Structures in Research Datasets and Their Ontic Implications lays the foundation for this special issue by discussing challenges of reusing datasets because of differences in their forms, structures, and the relationships between resources and their metadata. To find a solution to these challenges, the article looks at the relationship between metadata and characteristics of datasets from various domains, and shares findings on how metadata are used in different domains and contexts, and how the metadata schema impacts the discovery and reuse of the dataset.
Consortial Geospatial Data Collection: Toward Standards and Processes for Shared GeoBlacklight Metadata discusses metadata quality in a consortial geodata collection. Because datasets are contributed to a consortial collection by many institutions, the quality of metadata in any consortial database varies by contributing institution. The article examines metadata for geodata available in the GeoBlacklight, and proposes ways to enhance metadata that ultimately support discovery and access of geodata regardless of their discovery platform.
The third article, Open Metadata for Research Data Discovery in Canada, discusses a similar challenge. In order to create a central access point to research data across Canadian repositories, developers synchronized metadata across Canadian repositories of research data in a central, interdisciplinary discovery system called the Federated Research Data Repository (FRDR). Because FRDR pulls metadata from disciplinary repositories, a crosswalk to convert disciplinary metadata to a common metadata standard is crucial to its effectiveness. While reducing and transforming disciplinary metadata to elements that all metadata standards have in common can decrease the usefulness of the metadata for individual datasets, the project team was able to use facets to make datasets discoverable using more specialized fields.
Next, Describing Data: A Review of Metadata for Datasets in the Digital Commons Institutional Repository Platform: Problems and Recommendations looks at metadata across the Digital Commons platform, which has been adopted by many institutional data repositories. Since Digital Commons does not offer a standard metadata format for datasets, many institutions adapt Digital Commons' default metadata for their datasets. The article reviewed metadata for datasets hosted in Digital Commons repositories and confirmed the importance of having a standardized metadata for describing research data that work not only in individual Digital Commons instances, but also across the Digital Commons and outside of the repository.
These first four articles balance the need to develop specific metadata and granular discovery services appropriate to data in a particular discipline with the need to aggregate metadata in order to make the data discoverable in a service provider's environment or discoverable to people outside the discipline to support the reuse of those data. However, developing a domain-specific metadata standard for different types of data is still a challenge to many institutions. The final three articles are case studies from institutions that have developed metadata standards for specific types of data: weather data, research data, and historical topographic maps.
Modernizing Library Metadata for Historical Weather and Climate Data Collections describes a project where MARC records for the weather data in the collection of the National Center for Atmospheric Research (NCAR) Library were enhanced and converted to the ISO 19115 geospatial metadata standard in order to make the data more discoverable outside the Library's catalog. As many libraries' catalogs include datasets, the work outlined in the article will benefit libraries that consider providing separate discovery services for datasets together with other research data outside the library catalog.
Considerations and Challenges for Describing Historical Research Data: A Case Study describes a project to make a specific seismic dataset available in an institutional repository. The dataset was collected in 1970 and had very sparse documentation, so the authors worked extensively with one member of the team that collected the data to create textual descriptions about the context of the dataset. The repository software constrained the metadata format to Dublin Core, but these minimal metadata were supplemented by narrative descriptions. The article addresses the accessibility of complex data in repositories, proposing the possibility that data that is inaccessible to some (such as audio files to users with auditory disabilities) may expand access to others (such as audio files to users with visual disabilities).
Finally, Disciplinary Metadata in Libraries: A Case Study Applying the ISO 19115-North American Profile (NAP) for Describing Historical Topographic Maps describes a collaborative project among libraries in Ontario to inventory, digitize, georeference, and provide access to early topographic maps of the province. Once again, metadata are being combined from various libraries into a single discovery portal. The project used the ISO 19115 standard to describe these maps, resulting in robust access to a digitized map collection. The article includes the workflow of the project in detail for this important historical map collection.
Promoting discovery and reuse of data
As discussed in all of these articles, the balance between domain- or type-specific metadata and simpler metadata that can be easily aggregated into broader discovery and access services is an important consideration when designing metadata for data. In every case shared in the special issue, this balance was achieved in a different way, emphasizing the importance of the balancing process rather than the usefulness of a single standard that works well in every situation. However, common themes of the discussions are that the metadata should be supported by the community and promotes the discovery and reuse of the data. We hope that this special issue will provide great case studies for practitioners in technical services, public services, IT, as well as scholars to learn how different institutions and different domains have worked for describing and managing different types of data.
Lastly, we would like to thank the authors who contributed to this special issue. We also extend our gratitude to all reviewers and to Jung-Ran Park for the opportunity to edit the special issue.