Every data warehouse project has aspects that are new and untried, and slip-ups are natural. A penchant for burrowing in the mud unearths subtle problems near the beginning of the project. Identifying wrong solutions sooner, coupled with an aversion for obfuscation, leads to timely corrective action and saves much heartache later on.
Follow Your Nose
Like the physicist who studies the tracks left by the atom, the brog has a feel for what trails to follow and what data to stalk, iteratively gaining knowledge to further hone in on desired footprints. The business objectives are adjusted by the subtleties in the data, and the search for the right data is continually influenced by the business objectives.
Staffing an entire data warehouse project with brogs may be difficult. At a minimum, the project must be scoped appropriately and the data modeled correctly. That means that the lead business analyst—preferably, cum data modeler—should be a brog.
The analyst maintains the bird’s view by understanding the business objectives of the organization and elicits the questions the business wants answered to help achieve those objectives.
At the same time, with the help of the ETL designer, she digs into the source systems to discern the existence, accessibility and quality of the data needed to answer those questions. She keeps digging until reasonably confident of the feasibility of migrating the source data to support the business requirements.
A brog is ever practical. She starts with a data warehouse that does not do everything, but one that can mature to adulthood. Since she cannot foresee all that the data warehouse may face downstream, the analyst/modeler ensures the viability and extensibility of the warehouse by adhering to good data warehouse design and by balancing the exalted view of the business with the reality of mud-splattered data.
What secondary roles ought this analyst brog take on?
She can assist with ETL design. The ETL design and development phase of a data warehouse project consumes most of the effort and is burdened with most of the risks. Including the ETL designer in requirements gathering and the data audit, and actively drawing the analyst into the data staging effort, reduces errors and ramp-up time inherent in transition handoffs.
The analyst brog can also help specify an initial set of end-user report templates. She understands that, to the business users, the reports are the data warehouse. Regardless of the wizardry displayed in migrating the data, if the reporting application is difficult to use, or the response time is slow, the data warehouse has failed.
The analyst might also assist with data stewardship, protecting and assuring the integrity of the data as it migrates from the source systems to the data warehouse.
Where does one find a brog? Does this paragon exist?
Look for someone with an accumulated memory of transactional and dimensional systems. It is rare to find a person who has performed every role in both the transaction and data warehouse lifecycles, but it is possible to find someone who was actively involved in many aspects of both worlds.
And if you think you are a brog, try putting it on your resume. If nothing else, it should make for interesting conversation.
Esther Soloveichik is
senior consultant with Intrasphere Technologies, a technology consulting firm with a core focus on life sciences. Intrasphere provides end-to-end technology services and has successfully implemented large-scale projects for some of the world’s leading global companies, including Pfizer, Schering-Plough, Novartis and Eli Lilly, among others.