Analyzing Data › Basic Data Modeling Concepts

Basic Data Modeling Concepts

Data modeling is not an exact science. A business model is based largely on the pragmatic understanding of the business by the business practitioners. So, if two analysts study the same part of the business independently, it is likely that they develop models with differences.

In practice, data analysis proceeds in parallel with activity analysis to produce a business model where data supports the relevant business activities. This is described further in the chapter Analyzing Activities. Some representations of the underlying business reality are better than others, but successful data modeling depicts the business environment in a thorough and workable way that is generally understood and agreed on by all the business organizations involved. This means that group activities involving business staff in defining and confirming the model are vital to analysis.

It is necessary to achieve consensus among the group of business practitioners participating in the development project and often to coordinate with results achieved by other projects. In practice, it may not be possible or desirable to achieve perfect consistency at all times and throughout the whole organization. The business objectives of the project have to be taken into account. An early return on investment takes a higher priority than a demand for perfect consistency.

Bear in mind that data modeling is not database design. It does, provide the information that produce a database design. In fact, CA Gen transforms this automatically. The emphasis during analysis is on modeling business data, not on designing optimal data structures. Initially therefore, the model omits implementation details. Later, during system development, or sometimes in parallel with analysis if performance is already identified as an issue, the designer take steps to optimize the physical database implementation using all the entity volume and process frequency and entity type usage details that are collected in the business model. These volume and frequency details are collected as properties of entity types and elementary processes. Volume, frequency, and usage later is analyzed further, during distribution analysis, which is described in the chapter Analyzing Interactions.

The primary goal of data modeling is to depict accurately these fundamental elements of business information:

• Things with which the business deals
• Associations that exist between those things
• Detailed characteristics each thing possess

The following table summarizes the terms used for each concept at each level:

The following two levels of terminology are used for each of the three fundamental elements:

• Type or class level—A type or class is a description of some collection of fundamental elements that share similar traits.
• Occurrence or instance level—An occurrence or instance is a single instance of one of those elements. Data modeling is essentially a classification activity.

Those who are familiar with object-oriented concepts recognize the similarity to the distinction between object classes and objects or instances of object classes.