We have designed and maintained several large data warehouses for global enterprise level clients, as well as a multitude of smaller endeavors. This was a natural outcome from two fronts:
| 1. | Our historical function in statistical analysis and time series forecasting necessitated our developing expertise in the design and implementation of data warehouses. For us to be able to perform the stat work, 90% of the time we had to develop the underlying databases and to conjoin multiple disparate databases. In fact, it was this way that we initially developed our expertise in database design and administration, and |
| 2. | One of our underlying design requirements for all applications we develop is that the data structures we eventually develop are readily available and intuitively accessible to our clients for work and or analysis outside the scope of the initial requirements. This dictates that we at all times follow strict design criteria, and that our ultimate data architecture reflects the needs of the analyst. |
In most business environments, key data gets generated and stored at the level, and in the structure, of the corresponding business process. The information required to make a sale, schedule a delivery, etc. gets collected, and saved, in the form required to perform that specific event. The data structures generated in this manner frequently do not transfer readily to structures conducive to high level reporting and analysis. This remolding of the data is the true art of the data warehouse architect. Our extensive experience in exploratory data analysis gives us unique insights into the requirements of data for analytic support. This, coupled with our extensive experience in database design and implementation enables us to build highly accessible and useful data warehouses.