About BI and Data Modeling: Requirements and Data Modeling


There are two most common approaches to data modeling: relational modeling, often related to OLTP systems and dimensional modeling, which is a more appropriate technique for OLAP systems. Serra (2002) states that the process of building quality data models comes before the design of the first entity, starting with the understanding of the corporate model being addressed.

Collecting the business and data requirements is the foundation of the entire data warehouse effort—or at least it should be. Collecting the requirements is an art form, and it is one of the least natural activities for an IS organization. We give you techniques to make this job easier and hope to impress upon you the necessity of spending quality time on this step. (KIMBALL, 1998, p.6).

For any project aiming to build OLAP systems or OLTP systems, the reality with respect to the requirements is the same: they are the foundation of all the structure to be built and therefore, the necessary amount of time must be invested in order to prospect anything that is relevant. The more time invested in collecting and investigating requirements, the less time will be spent making unnecessary corrections in the future of the data model and the systems involved.

Fig. 2 - Definição de Requisitos e Modelagem Dimensional no Diagrama do Ciclo de Vida Dimensional do Negócio de Kimball
Definition of Requirements and Dimensional Modeling in the Kimball Business Dimensional Life-Cycle Diagram

From the business dimensional life-cycle model, on which Kimball’s (1998) methodology is based, note the importance not only of the process of defining the business requirements, but also, how much the dimensional data modeling process depends on these requirements, the dependency of the other processes of these two processes, and the critical path of the project, which tends to be strongly configured over this whole core set of processes.

KIMBALL (1998) states that in dimensional modeling, the definition of business requirements determines the data needed to meet the analytical requirements of business users. That is, for the ways of analysis that the business users intend to use can be made feasible, it is necessary to have the required data from the definition of the business requirements. A different approach than that used for the design of operational level systems is necessary to design data models that support such analyzes.

Even if there are applicability differences (and others) between data modeling techniques, relational or dimensional, the quality of the model will always be dependent on the quality of the requirements surveyed.

Text: Pedro Carneiro Jr.
Revision: Luis Cláudio R. da Silveira

These are the posts on the same “Enum and Quality in BI” monograph:

These will be the next posts on the same theme:

  • About BI and Data Modeling: Quality of Modeling, Data and Information
  • About BI and Data Modeling: Types of Data Modeling
    • About BI and Data Modeling: Relational Modeling
      • About BI and Data Modeling: Phases of Relational Data Modeling
      • About BI and Data Modeling: How to create an Entity-Relationship Diagram
    • About BI and Data Modeling: Dimensional Modeling
      • About BI and Data Modeling: Defining Granularity
      • About BI and Data Modeling: Detailing Dimensions
      • About BI and Data Modeling: Defining the Attributes of the Fact Table (s)
      • About BI and Data Modeling: Defining Aggregates


This short text is a mere Portuguese to English translation from part of my monograph “THE PERSISTENCE OF ENUMERATIONS IN POSTGRESQL DATABASES AND QUALITY IN BUSINESS INTELLIGENCE” (free translation of the title), also aliased as “Enum and Quality in BI”, which corresponds to a minor part of the document structure.


Image credits: