Experience

[Work in Progress]

Building a business glossary is not easy and creating standards for it isn't either. Applying standards will result in consistency and usability of the glossary data. So what standards need to be provided?  Standards need to be provided for each field however the standards will vary depending on organization and industry. Some standards we provided, the business definition needs to follow a language guideline such as American Psychological Association (APA) Style, Chicago Manual of Style or Modern Language Association (MLA) Style.

Standards

• Fields with Standards

  • Business Term Name

    • Title casing (APA guidelines)

  • Business Term Definition

    • Length

    • APA guidelines

  • Data Classification

    • Distinct types

  • Data Domain

  • Project Name

  • Acronym, Abbreviation, Synonym, Parent/Child Term, Related Term

  • Business Rule(s), Data Example(s)

Content Quality

• Standards

  • Standards will support content quality

• Content Reviews

  • Review of the business glossary will need to be carried out on a quarterly or annual basis (depending on volume)

• Identification

  • Issues found in the business glossary need to be addressed and a process to make the DG team aware of the issue is necessary

  • Utilize a submission method for users to submit issues found

[Work in Progress]

Data Governance Application

• Business term count

• Data asset count

• Number of users

Reference Data

• Reference term count

Data Governance Deliverables

• Number of deliverables

The Data Governance Artificial Intelligence framework is a focus on the collection of standards, rules, processes, responsibilities and security in the area of Artificial Intelligence.

Values

• Data Privacy

  • Personal data

  • Proprietary data

• Ethics

  • Bias

    • Identify and mitigate bias

    • Implement process to assess and track bias

  • Data Consent

• Model Transparency

  • Transparency provides openness and results in a minimization of unknowns regarding a model

  • A model needs to be transparent regardless of complexity

Standards

• Documentation

  • Project

    • Explanation of the object/process (i.e. model)

  • Workflows

Deliverable

• Data Requirements Template (DRT)

  • Define terms the business would need to know to understand the model's inputs/outputs and structure

• Data Mapping

  • Applicable if data flows aren't captured autonomously using the Data Governance application