Dimensional data modeling for data warehousing is a well-known technique and has been explained in great depth. These dimensional modeling techniques focus mainly on the data of the organization and how to organize it across multiple dimensions and facts. It is important to note, however, that the very existence of dimensional data modeling techniques is for better business alignment of the organization’s data with business requirements. This concept is currently working well and has been of use to the business for their analytical needs. The next step in the logical evolution of dimensional data modeling is to make it more intelligent by including relevant semantics into the data model. This article discusses a method of introducing semantics into the dimensional data model for a more realistic business alignment with the data present in the data warehouses.  

Embedding Semantics into the Data Model

 

Semantics in dimensional data model parlance is the act of introducing various business-level questions that the dimensional data model will answer. These questions can be based on the various permutations and combinations of the different dimensions and the levels in the hierarchy of these dimensions. A complete collection of these questions should be placed in a semantics query table and repeated for the other dimension combinations tables can be used to identify the information potential of dimensions and fact tables, as well as the information potential of the combinations of the dimension tables and their hierarchies.

 

Execution

 

Shown below is a series of steps that can be implemented to realize a semantics embedded data model.

  1. Create a semantics table for each of the dimensions of the dimensional data model and for the various permutations and combinations of the model.
  2. The structure of semantic embedded data models is shown in Figure 1 for a dimensional data model with four dimensions, such as product, customer, time and geography. Figure 2 shows the various combinations of semantics query tables.

 

 

 

  1. The “Query” column of the semantics tables should contain possible business queries, which the dimensional data model can answer.
  2. Create a complete set of possible business queries that a particular dimensional data model can answer and that can be obtained using set operators.

Benefits

 

The obvious benefit of embedding the semantic questions into a dimensional data model is the easy identification of various types of information that the data model can provide to an information seeker. This ensures that the information potential of a particular dimensional data model can be objectively defined or evaluated. Another benefit of embedding semantics into the data model is that a near total alignment of the business requirements with data warehouse (DW) data can be achieved. It can also enable the multilingual support to the underlying dimensional data model, thus enabling querying in multiple languages. Embedding semantics into the dimensional data model also will result in the optimization of the data model use. The complete use of a dimensional data model will also be comprehensible to even the regular business users who are not data modelers but understand the dimensional way of analyzing the data.

 

Further Work

 

The concept of the semantics embedded query tables can be further modified to associate the semantic queries with the underlying dimensional data model data in an automated fashion. This process ensures that only the semantic queries can be exposed to the user. When the user clicks the semantics queries, the underlying syntax queries are automatically built because of the association of the semantic queries to the underlying dimensional data model. This concept can be heavily leveraged in emerging search-enabled business intelligence (BI) technologies. This concept can also be further developed to enable information analysis at the semantic level by navigating across the various queries in these tables and by analyzing the information derived from these tables. We can then draw inferences regarding the kind of data or analysis that is possible from the underlying dimension data model for the information analysis done at the semantic level.

 

Dimensional data models can provide the necessary data, aligned per the business needs, in a syntactical manner. However, to derive business information from dimensional data models, its important to create queries using SQL. This is programming-intensive and is dependent on the skills of the BI professional doing the work. With a little innovation, however, semantics can be embedded into the dimensional data model, creating a semantically supportive data model. This will ensure that the dimensional data model is prepared to be completely in sync with business requirements, which can elevate the level of information delivery from the dimensional data models and the current syntactic level to a higher level of semantics. Syntactic dimensional data models will ensure intuitive use by all users and can be a powerful tool for self-service BI in the form of search-enabled BI.

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