Gold Layer Medallion Models

February 3, 2023

Data modeling is a crucial step in the development of a Gold Layer Medallion Data Warehouse Model. The two most common data modeling methodologies are the Kimball-style star schema-based data models and Inmon-style data marts.

Kimball-Style Star Schema-Based Data Models

The Kimball-style star schema data model is named after Ralph Kimball, a well-known data warehousing expert, and author. This data modeling style focuses on creating a clear and simple schema that accurately represents the analyzed business concepts. The star schema is centered around a central fact table, which holds the key metrics or measurements being analyzed and is joined to a series of dimension tables, which provide additional context and information about the metrics.

A typical example of a Kimball-style star schema-based data model is a retail sales analysis system. The central fact table would contain sales metrics such as revenue, units sold, and date of sale. Dimension tables could include information about customers, products, stores, and periods. By joining these dimension tables to the fact table, a user can perform analysis on retail sales data to answer questions like "What was the total revenue generated by a particular store over a specific period?" or "What was the best-selling product in a given region during the holiday season?"


Inmon-Style Data Marts

The Inmon-style data mart is named after Bill Inmon, another renowned data warehousing expert, and author. This data modeling style is centered around creating a series of specialized data marts that provide a focused view of specific business areas. Inmon-style data marts are typically built on top of an enterprise data warehouse and are designed to meet the specific needs of a particular department or business function.

An example of an Inmon-style data mart is a customer relationship management (CRM) system. The data mart would contain customer data, such as demographic information, purchase history, and customer interactions. This data mart would be designed to support the needs of the sales and marketing teams, providing them with the information they need to understand and interact with customers.



The choice between Kimball-style star schema-based data models and Inmon-style data marts depend on your organization's specific needs and goals. The right approach depends on various factors, including reporting requirements, project urgency, future staffing plans, frequency of changes, and the organization's culture.

Reporting requirements play a significant role in the decision-making process. If the reporting needs are broad and require integrated reporting, the Inmon approach is more suitable. On the other hand, if the reporting needs are more tactical and specific to business processes or teams, the Kimball approach is more appropriate.

The project's urgency is another factor to consider. The Inmon approach should be used if the organization has time to wait for the data warehouse's first delivery. However, if there is a tight deadline for the data warehouse, the Kimball approach is a better option.

The future staffing plan of the company also affects the decision. The Inmon method is more suitable if the company has the resources to maintain a large team of specialists. The Kimball approach is more appropriate if the team is expected to be small.

The frequency of changes and the stability of the reporting requirements and source systems are also essential factors. The Inmon approach is more flexible if the requirements and systems are expected to change rapidly. If the requirements and systems are stable, the Kimball method is suitable.

Finally, the organization's culture plays a role in the decision-making process. If the data warehouse's sponsors and managers understand its value proposition and are willing to accept long-lasting value from the investment, the Inmon approach is better. If the focus is on quick solutions for better reporting, the Kimball approach is better.

See how others are winning...
Learn why organizations are getting better outcomes using data and how Macula can help.
Start Now