Data modeling is a method that you can use for defining and analyzing which data is required within the scope of business process management in an organization. Data models define not only data elements, but also their structure and relationships. Data modeling techniques aim at modeling data in a standardized, consistent, and predictable manner so that it can be handled as a resource. Defining and observing data modeling standards is strongly recommended for all projects that require a standardized approach for defining and analyzing data within an organization.
Data modeling methods can be used for:
- managing data as a resource
- integrating information systems
- designing databases
Within the scope of data modeling you also have the possibility of detailing business requirements of a database. Therefore, data modeling is sometimes referred to as database modeling because a data model is eventually created in a database.
Data model types
In data modeling, there are three different model variants that can be applied to different phases of a project:
- Conceptual data models, also known as "domain models", are used to describe the semantics of a domain and to specify what facts can be expressed in a model. Here, a list of business objects or business terms like customer, purchase order, and employee is collected.
- Logical data models (LDMs) serve to describe the structure of a domain. They illustrate the logical entity types, their data attributes, as well as the relationships that exist between individual entity types.
- Physical data models (PDMs) describe how the internal schema of a database is designed. They illustrate data tables including their data columns, and the relationships between the tables.
Data modeling methods
Two methods are distinguished in data modeling:
- Bottom up: Button-up models are usually physical, application-specific models. They often result from reengineering efforts and are still incomplete from an enterprise point of view.
- Top down: Top-down models that are based on this modeling method are usually created in an abstract way by obtaining information from specialists acquainted with the subject area.
Data models created as a result of data modeling procedures are usually of the entity-relationship model (ERM) or UML class diagram type.
Start data modeling today
Data modeling is not just of interest for large enterprises, but it is a useful tool to structure requirements and to create a shared understanding of a domain. Start today with data modeling by downloading the free modeling tool ARIS Express.