Data modeling

Data modeling became popular as a means of identifying the data requirements and
ensuring data reuse. It provided the convention for the identification and definition
of every data element contained within the application, and offered a means of
documenting business relationships and definitions. The data model often became a
touchstone for understanding the meaning and context of corporate information.
There are a few important points to realize about data modeling:
A data model identifies data according to its business usage and meaning.
Because applications are unique in their business purpose, data models and their
content may differ across different systems, organizations and lines of business.
Data modeling is a logical construct, and arguably a business one. However, the term is
increasingly being used synonymously with physical database design, which represents
the design of the physical data structures usually as an outgrowth of the data model.
DM Review defines a logical data model as “a representation of business concepts laid
out in a visual format that clearly shows these concepts and their various relationships.”1
A data model is a documentation artifact that identifies and describes data using a
sanctioned convention—dimensional modeling and third-normal form are the two
prevailing ones—that ensures that data elements, their definition and their relationships
are clearly identified and distinguished.
Data modeling has also proved to be valuable as a means of documenting data
requirements. All too often systems are built based on processing needs with little
attention paid to the specific data requirements. Data modeling actually supports the
design process by ensuring that all data elements are clearly identified (using business
terminology), valued and described in business terms.
For instance, a data model might illustrate that an address will include a value called
state or province, denoted by a 2-character standard abbreviation. It would further
identify accepted values and if nulls were accepted. The benefit of this approach is
simple; the data content and rules are well-defined and documented. Irrespective of the
convention used—there are several different methods for developing and documenting
a data model—the value of creating a data model was that it represented the real-life
business relationships and definitions of data elements.