Episode 6: What is Metadata?

“Metadata.” It was one of those terms I kept hearing but didn’t fully understand. I’d seen Chike, the metadata analyst, working on complex lineage diagrams and data catalogs, and I knew metadata was important—but why? What exactly did it do? I knew it had something to do with “data about data,” but I didn’t fully grasp its importance.

That morning, Tola gave me my answer. “Ada,” she said, walking over with a confident smile, “it’s time you worked with metadata. I want you to work with Chike to document the data lineage for one of our operational reports. This will help you see how metadata keeps our data trustworthy and usable.”

With that, I grabbed my notebook and joined Chike at his desk, ready to get into the world of metadata.

Understanding Metadata

Chike started with a simple definition:

“Metadata is data about data,” he said. “It describes what the data is, where it’s from, how it’s used, and who can access it.”

He pulled up the bank’s data catalog, a centralized tool where all metadata was stored. “Think of this as our library catalog,” he explained. “It doesn’t contain the books themselves, but it tells us what books we have, where to find them, and what they’re about. Metadata works the same way for our data.”

He broke down metadata into specific types and how they apply to our work:

• Descriptive Metadata: Provides details about the content of the data, like titles, author names, creation dates, and descriptions. Example: The title of a report or survey name (e.g., "Monthly Operational Performance Report").
• Administrative Metadata: Information for managing the data, including who owns it, access permissions, and retention policies. Example: The HR team owns this dataset, and access is restricted to specific departments.
• Structural Metadata: Defines how data is organized and related to other data, like the relationship between tables in a database. Example: How tables in a database are linked, the hierarchical structure of a document, links between transaction logs and customer databases used in the report.
• Technical Metadata: Specifies the technical details of the data such as file formats, file sizes, storage location, and processing methods. Example: The report is a Power BI dashboard updated weekly from CSV data sources.
• Preservation Metadata: Documents efforts to maintain and preserve the dataset over time. Example: Backup schedules and archival methods for the data.
• Provenance Metadata: Tracks the history of data, including where it originated and how it was transformed. Example: Transaction data aggregated by branch and validated through ETL processes.
• Operational Metadata: Captures data usage and workflows. Example: This report is used by senior management to track branch performance.

Common Categories of Metadata

Chike paused to highlight how metadata is often grouped into broader categories for practical use:

• Business Metadata: Provides context to business users about the data’s meaning, purpose, and usage. Examples are dataset descriptions, business definitions, and usage policies. Helps non-technical users understand data’s relevance to their work.
• Technical Metadata: Focuses on the technical specifications and processes of the data. Examples are file formats, database schema, and ETL workflows. Enables IT teams to manage, integrate, and troubleshoot data systems.
• Operational Metadata: Tracks how data is used within workflows and processes. Examples are data update schedules, processing times, and user access logs. Helps organizations monitor data usage and optimize operational efficiency.

“I have to attend a workshop now,” Chike said, “when I get back and you are free, we will document the data lineage for the Monthly Operational Performance Report” and the metadata.

It has been an interesting one and so far, I have learnt that:

• Metadata is data about data: It describes content, structure, and management.
• There are different types of Metadata including Descriptive (what it is), structural (how it’s organized), and administrative (how it’s managed).
• Metadata is important: It enables data governance by making data findable, understandable, and reliable.

I certainly cannot wait to learn about data lineage and how metadata is actually documented. I would like to know though, how does metadata play a role in your work?

?