Business Intelligence - Data Warehousing in Detail


Many companies have a variety of information management systems. The management systems range in complexity from a single spreadsheet processed by a desktop computer to comprehensive client-server multi-user database applications running on an enterprise-wide network. The systems are designed to support functions of a particular business unit. These systems often have different data structures and platforms and thus do not lend themselves easily to efficient data sharing. To be effective for the whole organization, the information needs to be combined in a single data warehouse where it can be cross-referenced with data from other systems to provide meaningful analytics.

Our Data Warehouse methodology consists of the following:

Data Track
The definitions of the business requirements determine the data needed to address the business users' analytical requirements. Dimensional modeling concepts are used in this phase to design the optimal data schema for the Data Warehouse. This phase yields the logical data schema, the physical database and the importing of data from the various sources into the Data Warehouse

• Phase 1: Dimensional Modeling: is the name of a logical design technique often used for data warehouses. This method seeks to present the data in a standard framework that is intuitive and allows for high-performance access.



• Phase 2: Physical Design:The starting point for the physical data model design is the dimensional model developed in the first phase. This task also depends upon the outcome of the Technical Architecture design developed in the second track. A lot of the physical design is dependent upon software applications selected in this stage that will conform to the industry standards.



• Phase 3: Data Staging Design and Development: Data staging is one of the most important steps of the data warehouse. This is the stage where data is transformed from the source systems to the data marts identified through a series of ETL (Extract, Transform & Load) stages. A ten-step plan is followed for creating the data staging application for the data mart. Here too, industry standards will be conformed to in the choosing of applications and methodologies.

Technical Track

• Phase 1: Technical Architecture Design: The technical architecture covers the processes and tools we apply to the data. This section answers the question “how” – how do we get the data from its source to our data marts. It is made up of utilities, code and tools that bring the data warehouse to life.



• Phase 2: Product Selection & Installation:This phase involves evaluating and making recommendations for the hardware platform, DBMS platform, data staging tools, data access tools and the front-end application development platform. Industry standards will be conformed to in the selection process at this stage as well.

Application Track

• Phase 1: End User Application Specification: End user applications fill a critical gap in meeting the range of data access needs. Application specifications describe the report templates, user driven parameters and required calculations needed by the end users. These specifications ensure that the development team and business users have a common understanding of the applications to be delivered.
  The time and cost elements of the entire data warehouse project are greatly affected by the scope of the end user application.



• Phase 2: End User Application Design and Development:Following the drawing up of application specifications, the development of the end user applications involves configuring the tool metadata, application front-end and constructing the specified reports.

Deployment and Maintenance

• Phase 1: Deployment:Deployment represents the convergence of technology, data and end user applications accessible from the business user’s desktop. Extensive planning is required to ensure that the puzzle pieces fit together properly.



• Phase 2: Maintenance:The on-going task of maintaining the completed data warehouse is crucial. This phase involves developing a maintenance plan and a plan to address future growth of the warehouse.