How Portable Are the Metadata Standards for Scientific Data?


Published on

The one-covers-all approach in current metadata standards for scientific data has serious limitations in keeping up with the ever-growing data. This paper reports the findings from a survey to metadata standards in the scientific data domain and argues for the need for a metadata infrastructure. The survey collected 4400+ unique elements from 16 standards and categorized these elements into 9 categories. Findings from the data included that the highest counts of element occurred in the descriptive category and many of them overlapped with DC elements. This pattern also repeated in the elements co-occurred in different standards. A small number of semantically general elements appeared across the largest numbers of standards while the rest of the element co-occurrences formed a long tail with a wide range of specific semantics. The paper discussed implications of the findings in the context of metadata portability and infrastructure and pointed out that large, complex standards and widely varied naming practices are the major hurdles for building a metadata infrastructure.

Published in: Technology, Education
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

How Portable Are the Metadata Standards for Scientific Data?

  1. 1. How Portable Are the Metadata Standards for Scientific Data? Jian Qin Kai Li School of Information Studies Syracuse University Syracuse, NY, USA
  2. 2. Why study metadata portability? Complex, very large metadata standards •  are “…unwieldy to apply…” •  are “difficult to understand and enact in its entirety…” •  require customization to tailor to specific needs •  costly in time and personnel
  3. 3. Each standard has its own schema and tools… EPA Metadata Editor Metavist2 Morpho Protocol Registration SystemAVM Tagging Tool …that lead to duplicated efforts and interoperability problems
  4. 4. Metadata for scientific data is at the juncture of - Data-Driven Science RDF RDFa ORCID XML OWL DOI Technical Standards Infrastructure
  5. 5. A few big questions •  What action should and can we take at this juncture as a community of metadata practices? •  How much do we know about metadata standards for scientific data? •  How can we transform the current metadata standards into an infrastructure- driven service?
  6. 6. •  Portable •  Customizable •  Extendable •  Reusable •  Easy to use An infrastructure perspective for metadata
  7. 7. An attempt to define “metadata infrastructure” Semantically: • metadata elements, vocabularies, entities, and other metadata artifacts as the underlying foundation to build tools, software and applications Technically: • “a data model for describing the resources, aspects of metadata encoding and storage formats, metadata for web services, metadata tools, usage, modification, transformation, interoperability, and metadata crosswalk ” (CLARIN)
  8. 8. Portability is the key Building blocks of metadata Metadata for data citation Metadata for data discovery Metadata for data archiving Metadata for data quality Metadata for data provenance Metadata for data management Metadata generation output
  9. 9. How portable are metadata standards for scientific data? Two measures of metadata portability: • Co-occurrence of semantic elements: the times of semantically identical elements used in multiple standards • Degree of modularity: the degree of independence and self-descriptiveness of a sub-structure of concept/entity in metadata standards
  10. 10. Data •  5,800 elements from 16 scientific metadata standards Element Collection •  4,434 unique elements in terms of semantic Element De- duplication •  9 categories based on functionalities of the elements Categorization
  11. 11. Element distribution by standard NetCDF Climate and Forecast Metadata Convention (CF) 2427 elements Ecological Metadata Language
  12. 12. Element distribution by standard and category
  13. 13. Frequency of occurrences by category
  14. 14. Top occurring elements Element
  15. 15. Element co-occurrences across metadata standards Of 4,434 unique elements, 539 (12.16%) elements occurred in more than one standard Frequency of co-occurrences Numberofelements
  16. 16. Elements that most frequently co-occurred
  17. 17. Modularity Two levels of modularity: • Level 1: having multiple XML schema files for the whole standard; • Level 2: having separate schemas for entities such as person/organization, dataset, study, instrument, and subject Of the 6 standards with schema files, all of them belong to Level 1 modularity.
  18. 18. Discussion Portable metadata standards • Possible? • Feasible? • Advantages over the one-covers-all approach? A metadata infrastructure for scientific data • Bridge the gap between existing semantic and entity resources and metadata generation • Much to be researched…
  19. 19. Further research More questions than answers from this study: • What should a metadata infrastructure constitute? • How can the gaps be filled or narrowed between the infrastructure resources and metadata applications? • Is it possible or is there a need to streamline the metadata scheme design practice toward a metadata infrastructure? • …and the list can go on
  20. 20. Questions and comments?