The Dataspace methodology is derived from years of hands-on data warehousing experience combined with formal and on-site business training. It underlies each of our services packages and is a resource to every Dataspace consultant. The methodology consists of approximately 300 steps broken into six unique phases.
The Vision phase is to gain an understanding of the organization, its goals and the role that reporting systems will play in achieving those goals. During this phase we: Meet with key executives and users to understand their goals and information needs; Analyze the current reporting infrastructure; Define enterprise-wide reporting goals; Create a high-level enterprise data warehouse data model; and create a prototype of the proposed system. It is important to note that one of the issues we address during Vision is return on the warehousing investment.
In discovery, the user requirements for one warehouse functional area are fully ‘fleshed out’. The largest deliverable of the discovery phase is a user requirements report. This report covers topics such as: Primary, secondary, and ancillary target user communities; Security requirements; Required data elements; Anticipated usage patterns; and Logical data models. After management approval of the user requirements report, the architecture phase is initiated.
A data warehouse is not a static database but, rather, a complex system. It utilizes various data feeds, supports numerous users and tools, and must be prepared to recover from failure. Like any other system, the data warehouse must rest on a strong foundation. In many ways that foundation is the system’s architecture. During this phase we engineer that architecture. Some key deliverables of the architecture phase are: Physical database design and sizing, including security implementation design; Technical design report listing the programs that must be written; ETL and query tool selection; Refined workplan including target dates and resources required to meet those dates. It is important that the entire information systems organization, not just the data warehouse team, provide input into this step. This step defines how the data warehouse interacts with the company’s other systems, and ignoring the knowledge and input of the company’s experts can significantly undercut the design.
Click here for Specific Components of an Architecture
In the construction phase the system is actually constructed. Among this phase’s deliverables are: Programs and tools for refreshing, validating and archiving the data in the warehouse; Query tool and development metadata; Training plans and tools; Tools for administering user accounts and distributing software; Tools for analyzing performance and obtaining feedback from the user community.
In implementation, the system is moved to its production environment and distributed to users. Rollout usually takes place in stages, starting with a controlled pilot test and progressing to full distribution. After implementation, the company has: A populated data warehouse; Query software installed on user desktops; and Trained users.
The warehouse must be created iteratively. Users will not, themselves, fully understand their needs until they have interacted with the warehouse for a while. In addition, these needs are constantly shifting. Thus, the methodology includes an audit / iteration phase where the warehouse team utilizes the administrative tools created during implementation to verify that the system is meeting its goals. Modifications are made as necessary.