You are here: Visual LANSA Logical Modeler > 1. Introduction to Data Modeling > 1.1 What is Data Modeling? > 1.1.11 Approaches to Data Modeling

1.1.11 Approaches to Data Modeling

Two approaches are commonly used to build data models:

The top down approach begins with identifying the entities within the business. Entities are the major items which the users identify or work with. For each entity identified, create a list of attributes which make up that entity. Review the attributes to be sure that they cannot be broken into smaller pieces of data, i.e. they may be entities themselves. In this way, start at the highest level and fill in the details below.

The bottom up approach begins with identifying all the elements in the system.  This is a very long and difficult process of attempting to identify all the pieces of information the user comes in contact with. Once the elements are identified, they are grouped together to create entities.

The top down approach is recommended because it more closely relates to how the users define their business.  Following the top down approach may also prevent you from falling into the trap of physical modeling rather than logical modeling.

For example, a user may tell you that an order lists the product to be shipped to a customer.  There is no mention of an order line number.  An Architect might introduce the concept of an order line number simply because a key is needed for a database file.  This is physical modeling.