Continue in 2 seconds

How to Learn Data Modeling

Published
  • April 01 1999, 1:00am EST

The purpose of logical data modeling is to discover, analyze, define and standardize the discrete facts required by the business to conduct its activities. A data modeler must be concerned with correctly interpreting the information needs of the business and providing organization and reusability to the resulting data entities and attributes. To accomplish this purpose, a modeling candidate must learn certain skills, techniques and rules and apply those lessons rigorously and consistently. The best way to learn is a matter of some debate, but there are some proven methods for educating logical modelers that can be exploited and refined.

Many people learn modeling "on the job," some study a book or use computer-based education, others attend a public class, while still others learn by watching and imitating an experienced modeler perform what seems like magic. Ultimately, the best way to learn and retain the knowledge and skills of data modeling is a combination of all these methods. However, it is the order in which these methods are employed and the actual tools used that can separate a successful modeling educational experience from one that teaches the student very little or nothing.

Basics

Data modeling education does not assume much technical systems experience. Indeed, a novice can become an excellent logical data modeler with no programming or systems administration experience. Many data modeling professionals agree that programmers and systems administrators are not well suited to be data modelers since the skills needed for logical data modeling (i.e., abstract thinking, conceptual design, user liaison and communication) are different from the skills developed in the programming and systems administration fields (concrete thinking, hardware and software knowledge and systems integration experience). Basic requirements for a logical modeling candidate would include the proven ability to think abstractly and conceptually, to gather requirements from vague and often conflicting testimony, to demonstrate logical thought processes and to communicate well, including ­ especially ­ to listen well.

Communication skills are essential for the data modeler, even with the proliferation of documentation software, because much of the job of logical data modeling involves the translating and balancing of multiple user requirements and documenting the final results from the user perspective. A data modeler will usually present the completed model to an audience of users and database professionals and write clear, concise documentation to support models and craft appropriate definitions. Modelers also listen to the user(s) describe requirements and must cull the important and relevant facts from the discourse.

Most novice modelers come from the business community. By taking experienced business people as modeling trainees, a company can capitalize on their business knowledge to assist in "filling in the gaps" during modeling education. It is extremely helpful for a novice modeler to understand the business being discussed and analyzed and be able to concentrate on learning data modeling techniques.

Learning

A novice modeler should be assigned to a mentor. Unfortunately this is not always the case. Ideally this mentor should have experience, a formal modeling education and lived through "on the job" learning. If the mentor also understands how to be a good teacher, then the novice will get an even better education.

Many organizations are accustomed to training by experience ("on the job"), and to some degree this "education" is a proper way to teach data modeling and subsequent software development. However, most education from experience is best built on a foundation of knowledge, and with modeling most novices have no foundation upon which to build.

A better use of a novice's initial time and energy could be spent studying a computer-based education course teaching the elements of data and how understanding those elements in any business situation is "data modeling." The best example I know of interactive education in data modeling is Infostructor Pro from agpw, inc. Infostructor Pro allows a modeling analyst (novice or experienced) to follow the logical patterns of a modeling situation while learning the concepts of information gathering, analytical discernment and standardization, concluding with the development of models based on the cases presented throughout the "course." (Actually, the experienced modeler will probably get more out of this snappily-written gem from Bob Schmidt. The novice will only receive a first understanding of data modeling. Later, re- readings of Infostructor Pro by the novice will bear greater fruit as "on-the- job training" resonates in the real-life examples with which Schmidt loads his work.)

Unlike many computer-based courses, Infostructor Pro is presented in an engaging manner and provides opportunity for reinforced learning during the exercises. To be effective, computer-based courses in data modeling should, like Infostructor Pro, build on lessons learned and apply the techniques to case situations students would most likely encounter in the course of their modeling responsibilities.

Using the power and convenience of computer- based education is one of the better ways novice modelers can become exposed to the discipline of logical modeling; experiential training demonstrates methods for using that basic education to solve actual modeling problems. The case situations presented in a well-developed course form a progression of degrees of difficulty as well as ensure that the concepts of prior lessons are reinforced. For novice modelers with little or no data modeling experience, beginning their study with such a course is advisable.

However, any education taken in a vacuum does not fully reinforce the lessons taught in the course, making the time invested not complete. For example, each session of computer-based education should be followed by an actual modeling mini-project for the student led by the mentor and drawn from an actual modeling project for the organization. This activity has several purposes: 1) to provide concrete experiential learning of the concepts under study, 2) to demonstrate the business requirements that are being modeled and 3) to allow students to solve modeling problems commensurate with their skill levels and receive immediate results on their successful completion of the modeling assignment. The experiential learning exercise should be progressively difficult, and if possible, should occur in actual modeling sessions with users. This setting will also introduce the student to the facilitated sessions of data modeling discovery and will provide opportunity for the student to practice actual data requirements gathering with the mentor. (agpw, inc. enables this very feature with their exercise package that can be licensed in addition to the computer- based education.)

Most disciplines have formal courses that provide a student with the concepts and techniques, and data modeling is no different. There are several good workshops and seminars that can train a student in the concepts and techniques of data modeling. Modeling courses are designed to prepare the student to successfully gather requirements, execute the modeling techniques and document and discuss the models for users and technicians (programmers, DBA's and so on). These courses are most beneficial when they are coupled with the successful completion of computer-based education and mentor- driven performance.

Courses of varying intensity and knowledge transfer are offered through consulting companies (William G. Smith and Associates, Barnett Data Systems, Intelligent Solutions) and through training companies (Inteq Group, DPT Consulting and Training). Some universities, such as the University of Washington, offer distance learning in data management or offer data management courses as part of their IS curriculum. Do not confuse tool training with education in how to think about data clearly. Both are necessary, but the education in thinking about data and requirements is fundamental to any modeler, while tool training is appropriate for learning how to manipulate a particular piece of software.

Finding an appropriate course requires some diligence on the part of the student and mentor. To reduce confusion and assist in sorting through the details, the International Data Management Association (DAMA) has an educational services function, under which recognized vendors and instructors can display their course outlines and overviews of the programs in a central location. Although DAMA International does not endorse any specific course(s), providing this central location of data management instructional information delivers a service to the ever-expanding field of data management and data administration. To learn more about the DAMA educational services' "clearing house," visit the DAMA International Web site at http://www.dama.org. Many data administration/data management professionals find the affiliation with a professional organization such as DAMA to be very beneficial in their career and for continued education in this field.

Some excellent books published about modeling include A Practical Guide to Logical Data Modeling by George Tillman and Data Modeling by G. Lawrence Sanders. Magazines can offer the opportunity to refresh knowledge or start a learning process in a new area. You're reading this article in DM Review (http://www.dmreview.com); one strictly Web magazine, The Data Administration Newsletter (http://www.tdan.com), offers practical articles on data management in each quarterly edition.

As two of the most important steps in the development of information systems, logical data and process modeling require specific skills and knowledge that can come from a variety of sources and methods of learning. Each situation will dictate the path(s) the student will follow, but a multi-disciplined approach offers the best opportunity to acquire the knowledge and skills necessary for a successful data and process modeling career.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access