Silvermoon offers services to assist clients with the following aspects of Data Management.
- Dimensional Data Warehouse Design
- Business Data Modeling
- Business Data Requirements Analysis
Dimensional Data Warehouse Design
Among daily business operations, analytics have become of the highest importance. This is reinforced with the emergence of Big Data and its surrounding architectures, tools and roles, which require the formal identification of the data the enterprise wants to store and keep in a data warehouse in the long term. Moreover, analytical requirements have never been as dynamic and volatile as they are today.
Data warehouses are best developed via a “think big, start small” principle, rather than a “big bang” approach, in order to deliver business results more quickly.
Based on a solid enterprise business vocabulary defined during business requirements analysis on the one hand, and a Business Data Model on the other hand, standardized, more flexible and easier to deploy data warehouse “star schemas” can be designed using dimensional modeling techniques. We use Ralph Kimball’s design principles, including the concepts of Conformed Dimensions and Conformed Facts. By doing so, not only current analytical requirements can be supported, but also future requirements, without having to change the warehouse’s structure, reducing hereby its maintenance cost over time.
The dimensional data structures may become specialised data marts or the enterprise data warehouse itself. Whichever the case, the star schemas can easily be mapped to the source systems from which they are populated, hereby simplifying the Extract-Transform-Load (ETL) specifications. Moreover, their user-friendly structures are easy to deploy into downstream systems such as On Line Analytical Processing (OLAP) tools.
Business Data Modeling
Once the definition of business data requirements is correctly managed, these must be carefully analyzed and translated into a technology-independent enterprise Business Data Model which is the foundation for all “downstream” design models. These design models may represent a traditional production database, message definitions, an operational data store, a data warehouse or data marts. Business data modeling follows solid analysis rules, such as joint exhaustivity and mutual exclusivity, which are the source and mandatory condition for designing solid physical data models in an efficient, consistent and flexible way, thereby reducing the cost of future maintenance.
Using a Business Data Model provides an invaluable guide when integrating data stored in or sourced from multiple systems, including legacy data warehouses. At enterprise level, the Business Data Model is used as a Canonical Data Model to which those systems’ data structures and their content can be mapped. This facilitates integration, such as in the context of ETL design in Business Analytics projects when developing a data warehouse or data marts.
Business Data Requirements Analysis
Properly identifying and analyzing business requirements has always been a major challenge. Rather than being constrained by IT terminology, business people must be given the opportunity to express their requirements using their own business vocabulary in a structured manner.
We provide a precisely defined yet simple approach to identify and describe your business data requirements. This approach can be used to describe the business terms you use on a daily basis as well as to express your analytical requirements in a user-friendly way.
This results in higher quality of your business requirements and improved communication between Business and IT. Also, we find that our approach gives Business Experts more confidence in the resulting solution. Moreover, our approach stimulates ownership of requirements and data models and ultimately results in better data governance.
Last but not least, the definition of business data requirements must always take into account that business processes do not exist without business data, and vice versa. Hence the need to link business terms to the business processes that use them. This can be done by applying proven techniques, such as the concept of Bus Matrix for analytical requirements definition in Business Analytics projects.
For more information please contact Christian Palmaerts.
Email : christian at silvermoongroup.com