I just finished reading a Jim Harris blog “Chaos in the Big Data Brickyard”. When I saw the title, I thought that it must be a reference to the Indianapolis Motor Speedway, a.k.a.”The Brickyard”. I felt this was a reasonable assumption given the fact that NASCAR just ran their annual race there this past weekend. The interpretation of the title all depends on your frame of reference (which, in my case is usually formulated via ESPN). I assumed that somebody blew the race by miscalculating tire wear, fuel mileage, pit road speed or one of the myriad analytics used to manage the intricate strategies needed to win in the highly-competitive world of motorsports. To my surprise, that was not the case at all. The term “brickyard” taken without any context turned out to be another random brick of fact laid on an already crowded foundation.
Many have written about this topic before. And many will again. Context is what provides relevance to facts. Without a frame of reference into which a fact can be inserted it can easily become meaningless or, even worse, detrimental to the decision-making process.
Organizational context provided by business processes, reference data, master data, metadata all play critical roles in establishing the frame of reference needed to properly interpret the meaning and applicability of facts. As Jim describes, the rate at which facts are being created and shared is only going to increase. This creates even more room for error if we don’t establish the bounds within which those facts can (and should) be used. That means that we must have a comprehensive data management strategy in place if we are going to derive maximum value from data we intend to analyze.
As you look at analytics on demand, self-service BI, and big data analytics, it’s important to recognize that your data warehouse and master data management foundations are key components of an overarching data management strategy. These foundational elements must be designed based on high-value returns and must maintain a high-degree of flexibility to adapt to the agile demands of today’s data science approach to analytics. In his most recent blog,” A Good Data Warehouse Starts With a Firm Foundation“, my colleague and agile data warehouse expert, Brian Jones, put it as such: “Tackling the reference data integration issues up-front builds trust in the data as well as keeps the design emphasis focused on how the data will be used…”
Focusing on how the data will be used enables designers to focus on the high-value returns first and trusted data is essential to an efficient data-driven decision process. Creating a firm foundation refers to the reference data being the basis for analysis without over-engineering to the point of the foundation becoming brittle, which often becomes the case when you start with facts. Let’s face it bricks are hard and inflexible, not firm but adaptable.
So, in the future, I’ll consider my frame of reference before I jump to any conclusions about the title of a blog or article. No harm, no foul Jim! I consider myself more enlightened after reading your blog – as I often do. But, now I have to see who won Sunday’s race.
Join our upcoming webinar to learn how to reduce costs over a data warehouse life cycle by extending the model that was there on day one. Rather than rip-and-replace as business needs change, it is reuse-and-extend.