After the U.S. Office of Management and Governance issued the Open Data Policy, federal agencies set to the task of developing Enterprise Data Inventories to support a mandate of government transparency. The Department of the Interior (DOI) took this opportunity to create and implement a Metadata Management framework using an enterprise approach, properly documenting data for the greater understanding of it.

Some 90,000 employees reside across ten bureaus and eight agency offices of the DOI. Facilitating the discovery, access, and use of agency data by means of data provenance – meaning accuracy and quality – helps data become a sharable resource for users at every organizational level.

The DOI Metadata Implementation Guide, a framework for developing the metadata component for the agency’s data resource management, included instructions that all DOI bureaus or offices must deliver metadata records for their datasets with complete business, operational, and technical metadata.

“Business metadata categorizes the definition of business rules that apply to attribute properties.” That’s according to Raymond Obuch, an information technology specialist/data manager for the US Geological Survey (USGS) energy program who recently spoke at the DATAVERSITY® Enterprise Data World Conference. His presentation was titled: Department of the Interior Metadata Implementation Guide—Framework for Developing the Metadata Component for Data Resource Management. Business metadata, Obuch said, is aimed at relating the business perspective to the metadata user – the business professional.

Operational metadata is administrative metadata. “It’s about data file size, time of last load, references to procedural scripts,” and other lifecycle attributes of a resource such as acquisition and lineage, Obuch noted.

Technical metadata is a database’s physical characteristics – the actual definitions of fields, such as table and column names, in a database system.

Maintaining a bureau or office metadata catalog with metadata tags was meant to facilitate data community delineation; ease advanced metadata search for data discovery, data geographic location, and digital object identifiers; and ensure minimum metadata requirements. These and other big benefits are the outcome of data integration where sources of data are pulled in and cataloged, Obuch said.

“There is lots of good data out there and a lot of portals to help you discover it, but you have to be able to make sense of it to use it,” cautioned Obuch, who served as representative to the DOI Data Advisory Committee, a committee that works to promote data sharing across DOI and to ensure data standardization and data interoperability. Focusing on metadata tags for search and a robust data dictionary for use has been a priority.

Phasing the Metadata Implementation

The DOI came up with a phased approach to metadata implementation across its various bureaus and offices, Obuch told the audience, accounting for different maturity levels. “Each group implements actions based on where they are,” he said.

The three phases of metadata implementation encompassed:

• A “getting started” plan of activities required to establish the foundational aspects and requirements for successful bureau or office Metadata Management.
• An “implementation and management metadata” effort that built on the planning actions by implementing processes and workflows for metadata creation, management, and publication that could become the framework for long-term Metadata Management.
• An “improving Metadata Management” scheme for providing a consistent approach to metadata evaluation, metadata quality, and Data Architecture for data hosting, data services, a metadata catalog, and overall Metadata Management across all functional areas.

Structuring the Phases

The various phases, Obuch said, account in various ways (depending on organizational maturity) for:

• Key Drivers: The reasons why an action needs to occur. Overall, the key driver is creating a single trusted source for organizational data to enhance a user’s ability to search, find, manage, and use bureau or office authoritative data on a routine basis.
• Description: The steps and processes that support completion of the action. In Phase 1, this included performing a baseline inventory of data and metadata holdings, including digital and paper materials, to facilitate the identification and prioritization of next steps in metadata implementation and management.
• Dependencies: The people, processes, and funding that are required for the success of an action. In Phase 3, this included personnel who have technical expertise in installing, configuring, and running metadata catalog software and the integration of metadata management into business processes and functions.
• Responsible Roles: The work and associated responsibilities that need to be assigned to individuals to complete the action. For example, as part of Phase 3, business analysts, data modelers, subject matter experts and data stewards created and maintained the business metadata while data stewards and program staff created and maintained technical and operational metadata.
• Deliverables: The expected outcomes and products resulting from the action. In Phase 2, the deliverable was a current DOI Metadata Catalog with each metadata record flagged as public or nonpublic. 
• Measurements: The metrics that are used to measure the success of an action, usually tied to deliverables. In Phase 2, that could include the level of adoption and use of these tools across the bureau and office.
• Recommended Methods and Tools: The references for metadata implementation, best practices, and case studies in Phase 3.

“First create and maintain metadata,” Obuch said. “Then publish metadata to one source.”

Want to learn more about DATAVERSITY’s upcoming events? Check out our current lineup of online and face-to-face conferences here.

Image used under license from Shutterstock.com