As a graphical representation of the information requirements for a given business area, a logical data model is constructed by taking the data descriptions depicted in a conceptual data model and introducing associated elements, definitions and greater context for the data’s structure.
This stage is important because while the more streamlined conceptual data model is more easily communicated, the lack of context can make it difficult to move from modeling to implementation. More detail is required to support that progression. Such detail includes defining the owned attributes, primary keys, foreign keys, relationship cardinality and describing entities and classes. At this stage, the nature of relationships between data is established and defined, and data from different systems is normalized.
The three different types of data models provide increasing degrees of context and detail, so we can view their use sequentially. Therefore, a logical data model should be considered once the conceptual data model has been built.
This more structured stage of data modeling is most relevant during application design, when it can serve as a communication mechanism in the more technical environments where database analysts and designers work. It helps us understand the details of the data to a greater degree than conceptual data models – but similarly stops short of providing perspective on how it should be implemented.
As with conceptual data modeling, this means teams aren’t bound to technological considerations. This is important as the nature of technology in organizations is often dynamic.