In any data warehouse implementation there are many different considerations which should in place before the final physical setting up. This is to avoid in problems related to quality of data and consistencies in data processes.
A conceptual schema is an abstract definition of the whole project. In the case of data warehouse and business intelligence system, a conceptual schema represents the map of concepts and their relationships. A data warehouse is built upon a business data architecture and the business data architecture defined all the common business structures that pertain to the overall activities of the business enterprise.
The conceptual schema describes the semantics of a company. It represents the series of assertions and rules pertaining to the nature of the business processes, entities and events. In particular, the conceptual schema the thing which are very significant to the company, a term called entity classes and the characteristics of these things, a term called attributes. The association between pairs of the those things of significance is called a relationship.
A conceptual schema, although it greatly represents the data warehouse and the common structure of data, is not a database design. It exists as different levels of abstractions. These abstractions are the basis for the implementation of a physical database.
Any conceptual schema is done by using a human oriented natural language. This natural language is used to fine elementary facts. The conceptual schema is totally independent of any implementation whether database or non-IT implementation.
The data model and query design of a business architecture and should be performed at the conceptual level and then mapped to other levels. This means that at the conceptual schema everything should be gotten right in the first place. And then as the business grows and evolves changes can be made later. But many keen data architects usually design the data model for scalability. This means that all business growth and evolution are already taken into consideration in the concept schema level.
A conceptual schema should follow the following criteria: expressability, clarity, simplicity and orthogonality, semantic stability, semantic relevance, good validation mechanisms, abstraction mechanisms and formal foundation.
Making the conceptual schema commonly involves the close coordination between the domain expert and data modeler. The domain expert best understands the application domain. He or she understands the scope of the enterprise activities including the individual roles of the staff and the clients. He or she also understands the scope of the products and services involved. On the other hand, the task of the data modeler is to formalize the informal knowledge of the domain expert.
As the case should be, the communication between the domain expert and the data modeler involves verbalizing fact instances from data use cases, verbalizing fact types in natural language, validating rules in natural language and validating rules using sample populations.
With close coordination between the domain expert and the data modeler, the expected output should be a conceptual schema that have data expressed as elementary facts in plain English sentences (or in any language appropriate depending on the users). Facts are also laid out on how they are group into structures.
Conceptual schema is really intensively used not just in database implementation in many other IT systems as well. It is plain definition of abstract ideas and entities whereby all technical specifications are taken from. Even in other fields not related to IT, having a conceptual schema before actual implementation of a plan helps facilitate a project smoothly and efficiently.