Metadata – The Data Quality Prerequisite

Think about these questions.

• What data do I have?

• What is it called?

• What does it mean?

• What is its type?

• Where did it come from?

• Where does it live?

• What happened to it along the way?

• How did it get there?

• How do I access it?

• What is its accepted data quality?

• Who do I call if I have a question?

Can you answer them for the data you’re using on a regular basis?  What about for augmenting data sets?  Creating new datasets?  New reports or dashboards?  AI models?

When I talk to clients about what they really worry about, the answer usually traces back to ‘data quality’ (I’ll write about that more next week).  My response always comes back to being able to answer simple ‘metadata’ questions.  I consider those the minimum level of metadata that should be available to data consumers.

If you can’t answer those simple questions, ‘data quality’ becomes anecdotal at best.  The result is a lack of trust in data outputs (including AI) and, more importantly, the decisions or actions taken from that output.

Most everyone will agree that Data Governance is a prerequisite for AI Governance.  In that same vein, metadata is a prerequisite for understanding and taking action on ‘data quality.’