Learn how to build the most optimal data models for your reports by attending one of our courses

 We have created a training course on data models, teaching you how to solve business scenarios in Power BI, Power Pivot and Analysis Services by adjusting your data model. Your DAX code becomes easier, faster and more robust.

Data Modeling vs. Requirements Modeling

To understand data modeling and how it contributes to software engineering, it is best to take a step back and examine how data modeling fits into requirements modeling. 

Requirements modeling in software engineering is part of analysis and design. It's the planning stage of developing a software application. Requirements modeling focuses on the 'what', not the 'how'. In other words, requirements modeling identifies what requirements the application must meet in order to be successful. 

Requirements modeling includes many sub-stages, one of them being data modeling. Generally speaking, requirements modeling will begin with scenario-based modeling, which results in creating a use case. A use case, simply put, is a primary example of how the software application will be used and what it is expected to do. Once a use case exists, one of the stages that follow will be data modeling. 

Conceptual, Logical, and Physical Modeling

Traditionally, data modeling is comprised of three stages: conceptual, logical, and physical modeling. These stages begin at the most abstract, or general, representation of the data (conceptual modeling). Then they proceed systematically to the most detailed representation using physical modeling. 

Conceptual modeling primarily identifies the entities (data objects) that the application will deal with, and traditionally does not describe the attributes of those entities. This stage is deliberately abstract, to avoid becoming 'bogged down in details'. This way it allows all of the stakeholders to participate at this stage, without requiring them to have any specific technical knowledge. For example, the representatives of the company who have commissioned the application (who are not necessarily technically-minded) can contribute meaningfully alongside the software engineers (who are necessarily technically-minded) in designing the application. 

Logical modeling uses the entities identified during the conceptual modeling stage, and then identifies the attributes these entities should have, along with the relationships between the entities. 

Physical modeling is the most detailed stage of data modeling. The previously defined entities, attributes, and relationships are used to design the database structure, including all of the necessary details (like the number of tables, columns, and data types needed).