Data Warehouse design and development typically follows a basic, common path. First, you gather requirements and create some form of the bus matrix as a deliverable which documents the requirements of the upcoming dimensional modeling work that needs completed. Second, you must design your dimensional models. The dimensional models for a data warehouse consist of a relational data warehouse and quite likely an OLAP model as well (cube design). It is of utmost importance that your dimensional models be as complete as possible. Several inputs need to be taken into account when creating dimensional models including but not limited to: 1) Business Requirements 2) Performance Requirements 3) Extract, Transform, and Load (ETL) Requirements 4) Data Lineage If you proceed into further DW/BI tasks and then have to revisit your models at a later time you will incur retooling your ETL processes and BI applications. Take the time to work closely with both your clients and your DW/BI architect to ensure your dimensional models are as complete as possible. Happy Modeling
