News Stay informed about the latest enterprise technology news and product updates.

Data modeling and unstructured data

The data model has long been recognized as one of the essentials in the structured environment, but do we need a data model for the unstructured environment as well?

This article originally appeared on the BeyeNETWORK.

The data model has long been recognized as one of the essentials in the structured environment. The data model is useful to designers, to developers and to the people who do maintenance after the system is built. The data model also has ancillary uses as a communications device for conversations with management, end users, contractors and so forth.


The data model serves as an intellectual road map for both current and future development. Without a data model, it would be almost impossible to build an architecture with continuity over time. It is through the data model that developers know that their project will fit with other development projects in the future.

But all of this foundation for a data model is for the structured environment. An interesting question arises – do we need a data model for the unstructured environment as well?

In order to answer this question, we have to take a look at the fabric of the data model in the structured environment. In the structured environment, the data model consists entirely of pure, plain and simple metadata.

Now we go to the unstructured environment and start looking for metadata and what do we find? We find business terms, not the classical “data about data.” In the unstructured environment, we find business terms such as “account,” “customer,” “payment,” “terms,” “interest rate,” “due date,” “final payment” and so forth. Those same terms are found in the structured data model, but in the structured data model, those terms exist with a great deal of formalism. The terms in the structured data model carry other connotations such as key, index, attribute, foreign key and so forth.

So there are some similarities and some differences between what is found in the structured environment and what is found in the unstructured environment.

With that understanding, can an unstructured data model be constructed from business terms? And if you construct such a model, what are you left with? The answer is that if you construct an unstructured data model using business terms, you end up with a business model. The business model is much more a description of the primitives of the business than anything else.

Like a data model, a business model defines only the most basic elements of the business. Derived data is removed from the data model, and so derived business information is also removed from the business model. In doing so, the business model only reflects the most visceral aspects of the business.

So who might use a business model? The answer is almost anyone. The business model can be used by management, by accountants, by salespeople, by marketing people and so forth. The business model becomes the basis for clear communications and a means by which the unstructured data starts to achieve some semblance of uniformity.

But, ironically, the business model does not become a basis for development in the unstructured environment. Development in the unstructured environment happens in an entirely different manner than development in the structured environment.

Another difference between the structured and unstructured data models is in their formalism. The structured data model is formal while the unstructured data model/business model is informal. This difference is due to the fundamental differences between the environments. And, the structured data model has as its primary users developers and systems personnel. The unstructured environment has as its users businesspeople.

The data model/business model in the unstructured environment is an interesting construct.

Dig Deeper on Data modeling tools and techniques