Data models visually represent the concepts that are important to the business, the information associated with them and the relationships between them. The two most common types of data models found in the BABOK® Guide that are used by business analysts are the Entity-Relationship Diagram (ERD) and the class diagram. ERDs tend to be selected on projects where a relational database is part of the solution, while class diagrams are a better fit for object-oriented development efforts. The BABOK® Guide considers building data models is one of the 16 required techniques in the fundamental knowledge base of an effective business analyst.
Logical data models look at the concepts, attributes and relationships for the information relevant to the organization in detail or at a high level. Physical data models describe how data is stored and managed by a software application that is part of the solution scope. Concepts are something significant to the organization about which the organization needs data. Attributes are used to define specific pieces of information associated with a concept, such as its name, acceptable values and a description. Relationships are significant business associations between concepts.
In addition to defining the information and its relationships, data models also describe the context, use and validity of their business information using metadata, or “data about data”. Metadata tells the business analyst when and why information in a system is being changed in some way.
Data models provide you with a strong vehicle for transitioning through the planning, analysis, design and implementation of a project. They possess rigorous rules for correctness and completeness, which typically results in a more accurate final product. In some cases, stakeholders can find detailed data models difficult to understand, so you must be prepared to build them to the appropriate level of detail and be able to explain them thoroughly to their audience if required.