News Stay informed about the latest enterprise technology news and product updates.

Business data models: Building a meta data strategy

This model contains four critical subject areas that should to be addressed by any meta data model.

This article originally appeared on the BeyeNETWORK.

This model contains four critical subject areas that should to be addressed by any meta data model:

  • Logical Subject and Business Area Environment – this portion of the model describes the business. It consists of the business data model and other meta data providing the foundation for the business intelligence environment.
  • Data Quality Environment – this portion of the model contains the quality expectations and the on-going information about the quality being processed. This area will become increasingly important as companies apply meta data in support of Sarbanes-Oxley and other regulations.
  • Physical Programming Environment – this portion of the model describes the physical tables and columns and the processes used to move data through the business intelligence environment.
  • Security Environment – this portion of the model describes usage authority as well as the actual access to the data in the business intelligence environment.

The ETL vendors have recognized the importance of meta data, and modern ETL tools contain a meta data repository, the mechanisms to import data into the repository from other tools (such as the modeling tools), and other facilities to export the meta data into the end-user access tools. This article addresses the first part of the complete model, which is the logical subject and business area environment model. Entities within this model would include:

  • Subject Area - A Subject Area is a major category of data relevant to the business, typically used across one or more Business Areas. Examples of Subject Areas are Customers, Products, and Organizations.
  • Business Area - A Business Area is a functional grouping of similar business processes and activities of interest to a Data Steward. Examples are Finance, Human Resources, and Marketing.
  • Data Steward - A Data Steward is a person or group of people responsible for defining, setting, and enforcing the policies for data. Data Stewards generally represent a business community.
  • Entity - An Entity is a person, place, thing or event about which the company records information. More specifically, in database analysis and design, an Entity is a grouping of related information.
  • Attribute - An Attribute represents a single discrete item of information or a property of the entity. The attribute may be a key attribute, which uniquely identifies an entity, or it may be a non-key attribute, which is not used (in whole or in part) to uniquely identify an entity. Examples are Customer ID, Medical Claim Reason Code, Automobile Manufacture Date, and Product Wholesale Cost.

As we review this model, we see that a subject area can be related to multiple business areas and that each business area has a data steward to represent it. A subject area also may have many entities with each entity also having many attributes.

Here is a copy of this portion of the meta data model:

This is a starter model. As you apply it to your environment, you may need to include additional entities and attributes to make it relevant. To ensure that the content is understood, provide the meta data for the meta data model in the form of definitions for each of the entities and attributes that are included.

In summary, take what works from our model and make it yours. Use it to compare models presented by the vendors or to create your own robust meta data architecture.

Dig Deeper on Business intelligence architecture and integration

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.