Data characteristics are defined during data modeling, a process where a data model is created by applying a data model theory in order to create a data model instance. A data model theory is a formal description of a data model.
Data modeling is also a process of structuring data and organizing data so that the data structures will then be the basis for the implementation of a database management system. Also in addition to organizing and defining data, the data modeling process also implicitly or explicitly imposes limitation and constraints on the data within the structure.
Data models are based on business rules. This is because business rules are abstracts and intangible concept that the database management system, which is basically a computer system, cannot understand. Business rules converted to data models convert data into format that the computer will finally be able to understand and thus implement.
Here is an example of a draft business rule that will be the basis of a data model:
A. The aspect being modeled is a Product Line
B. The things of interest, herein referred to as "Things" include
(2) Product Categories
(3) Product Characteristics
C. These "Things" are related as follows:
(1) A product can be in one and only one Product_Category;
(2) a product can have zero, one or many Product_Characteristics.
D. The other characteristics of the "Things" include
(1) A product which can have either zero or one typical_buying_price and
(2) A product can have either zero or one typical_selling_price.
E. Sample data may include products to be determined by the company
F. Typical inquiries may include typically selling price for a certain number of products
It is apparent from the draft business rule that data characteristics are everywhere. As mentioned earlier, data characteristics can either be developed directly through measurement or indirectly through derivation, from a feature of an object or event.
For instance, if a product is a T-shirt, the data characteristics that are developed by direct measurement are
(1) material composition of the T-shirt,
(2) size range of available T-shirts,
(3) style of the T-shirt and
(4) supplier of the T-shirts.
On the other hand, some of the data characteristics which can be developed through derivation may include
(1) bulk price of the T-shirts which can be derived depending on the number of orders and
(2) shipment price of the T-shirts which also depends on the number of orders.
Data characteristics are very important in an area of data modeling called entity-relationship model (ERM). The ERM is a representation of structured data where final product is an entity-relationship diagram (ERD).
From the data models where the data characteristics are defined, relationships among the data, or more technically known as "entities" defined. Data characteristics are also known as "attributes" in data modeling jargon. Relationships could be as simple as "An employee entity may have a attribute which a social security number.
There are various types of relationships in ERM. Some may have many to one, one to many, many to many or one to one. Database implementers need to give ERM design a very careful consideration because any slight failure can result in weak data integrity and the resulting flaw could be hard to trace. As recommended by database experts, it is always good to draft a database plan using plain English language and data characteristics should also be defined likewise.