The short answer is “yes.” The need to focus on business process measurement events, plus grain, dimensions and facts, is as important as ever. When dimensional modeling was popularized several decades ago, we concentrated on schema designs that delivered query performance and ease-of-use. But as I described in Design Tip #160 Think Dimensionally (Beyond Data […]
Is Dimensional Modeling Still Relevant?
The short answer is “yes.” The need to focus on business process measurement events, plus grain, dimensions and facts, is as important as ever.
When dimensional modeling was popularized several decades ago, we concentrated on schema designs that delivered query performance and ease-of-use. But as I described in Design Tip #160 Think Dimensionally (Beyond Data Modeling), focusing on business process events and metrics, along with the associated descriptive context, is useful in other situations including:
- Synthesizing business requirements (see Design Tip #3 Focus on Business Processes, Not Business Departments!)
- Defining and prioritizing manageable project scope (see The Bottom-Up Misnomer)
- Designing the overall DW/BI data architecture (see The Matrix: Revisited)
- Establishing data stewardship/governance responsibilities
- Presenting a business view of the organization’s information
“Thinking dimensionally” is vital, regardless of your underlying technology and data distribution strategies.
Meanwhile, technology has certainly advanced since dimensional models were introduced in the 1980s. Between NoSQL, virtualization and visualization technologies, some have predicted that the dimensional model must be on its deathbed. While some organizations are migrating some data to new platforms, relational and multidimensional solutions are far from obsolete. Dimensional schemas will continue in this space (and contrary to several big data pundits, they’ll continue to successfully handle atomic, granular data just as they have for decades.) Non-relational platforms are providing relational support. And dimensional models will help ensure consistent results from non-relational solutions.
Dimensional models organize data in a way that makes sense to business people, albeit potentially virtually. Some virtualization tools contend there’s no need to physically populate dimensional models. If your environment can deliver satisfactory query performance by accessing disparate data via a virtual logical dimensional model that provides the business view, perhaps you should give it a try. (I’d also recommend you conduct serious performance tests, evaluate the costs of ownership, think about the users’ requirements for slowly changing descriptive data, and so on.)
Some visualization tools contend there’s no need to dimensionalize because the tools prefer a flattened many-columned “big wide table.” In this case, the dimensional framework is beneficial for managing your data assets so consistent attribute-rich master dimension data is repeatedly attached to business process performance metrics rather than repeatedly embedding potentially inconsistent attributes in flattened data sets scattered around the organization. You should govern your dimension assets before presenting data per the visualization tool’s preferences. This will also help “future proof,” just in case the next great thing comes along with a new set of tool-specific requirements.
Finally, some dismiss the dimensional model as old news simply because it’s been a generally-accepted best practice for several decades. This one is hard to argue with. Yes, I’ve been doing dimensional modeling for over 30 years. Ralph Kimball’s original Data Warehouse Toolkit was published in 1996; I co-authored updated editions with him in 2002 and 2013. The concepts have withstood the test of time! Dimensional modeling has helped countless organizations across every industry make better business decisions which should be the true measure of DW/BI success. While it’s fun to try something new, why dismiss a proven, valuable technique that has provided positive return on investment?
Dimensional modeling doesn’t garner the vendor mindshare that big data and other technologies currently do, but that doesn’t mean it’s no longer relevant. There will most certainly be complementary roles in the DW/BI space for the foreseeable future.
Time marches on and soon the collective retirement of the Kimball Group will be upon us. In my final Design Tip, I would like to share the perspective for DW/BI success I’ve gained from my 26 years in the data warehouse/business intelligence industry. While data warehousing has been around now for a long while, there […]
For my final Design Tip, I’m returning to a fundamental theme that’s not rocket science, but too often ignored: business-IT collaboration. If you buy into the proposition that the true measure of DW/BI success is business acceptance of the deliverables to improve their decision-making, then buying into the importance of collaboration should be easy. Achieving […]
Countless organizations have created mature dimensional data warehouses that are considered tremendous successes within their organizations. These data warehouse environments support key reporting and analysis requirements for the enterprise. Many are capable of supporting self-serve data access and analysis capabilities for disparate business users. Nonetheless, regardless of the success achieved by these dimensional data warehouses, […]
Over the years, we’ve worked with countless exemplary DW/BI project team members: smart, skilled, dedicated, and motivated, coupled with a healthy dose of mutual trust, respect, and camaraderie with their teammates. Teams with members who possess these characteristics tend to fire on all cylinders, with the resulting whole often greater than the sum of the […]
In most cases, metadata is a neglected area of the DW/BI system; however, an increasing number of DW/BI teams have made positive strides in delivering business metadata to their users. This Design Tip looks beyond the business metadata to suggest several opportunities for leveraging ETL process metadata to improve data warehouse operations. The goal is […]
There are three fundamental types of fact tables in the data warehouse presentation area: transaction fact tables, periodic snapshot fact tables, and accumulating snapshot fact tables. Most DW/BI design teams are very familiar with transaction fact tables. They are the most common fact table type and are often the primary workhorse schema for many organizations. […]
Dimensional designs often need to accommodate multivalued dimensions. Patients can have multiple diagnoses. Students can have multiple majors. Consumers can have multiple hobbies or interests. Commercial customers can have multiple industry classifications. Employees can have multiple skills or certifications. Products can have multiple optional features. Bank accounts can have multiple customers. The multivalued dimension challenge […]
We’re sad to tell you that Warren passed away yesterday. He was diagnosed with an aggressive glioblastoma multiforme brain tumor nearly 14 months ago. Warren and his wife Elizabeth fought the disease with incredible determination and optimism while remaining focused on living life to the fullest. All of us at the Kimball Group will miss […]
Regular readers know we stress the importance of focusing on business requirements when designing dimensional data models to support the data warehouse/business intelligence (DW/BI) environment. As described in Design Tip #157, it is critical to include business partners in the dimensional design process. But including business representatives on the design team obviously increases the size […]