How to Build an Enterprise-Level Data Analytics Framework

Video produced by Steve Nathans-Kelly

At Data Summit Connect Fall 2020, John O'Brien, CEO and principal advisor at Radiant Advisors, outlined the components of a modern data architecture.

Picture of John O'BrienAs a conceptual architecture, said O'Brien, it is important to ensure that we are anchored in architecture, and, in terms of architecture priorities, it is necessary to ask, What is the analytics capability we're trying to deliver in the company? "This means it is a little different than a project," he said. "You need to look at the analytics capabilities involved."

Considering the conceptual architecture, O'Brien stated, "In our world, what we see and believe is that business intelligence, data warehousing, reporting, and OLAP is not going away. It serves a purpose within the business that is what I consider operational performance management. As a company that is operating and that is executing, you set goals, you have metrics, you work to achieve those goals and running the company hasn't changed, nor is it going away."

Additionally, there is a new area that Radiant calls 'enterprise self-service data analytics,' or 'business data enablement,' which, O'Brien said, means that we're looking at enabling the business to work with data. The business is going to have to do discovery, gain insights, and validate things that help a data warehouse project, but also take care of themselves and maybe other teams don't need to get involved. It doesn't need to be a project per se. So we want to focus on that capability because in itself it adds a lot of value to the business, as well.

The fact that everybody's able to connect to data, find data, trust data, follow a modern analytics lifecycle, which is one of our methodologies, means that you have increased productivity in working with data. "And then our third key spectrum, if you will, if the left-hand side follows the classic industry terminology of saying descriptive and diagnostic analytics—saying looking at what's happened and understanding how that might've happened—then the right hand side is more into the predictive analytics and prescriptive analytics world where we shift from 'Here's the data and what happened yesterday'  to 'Here's what we think is going to happen today, next month, in the next minute.' And so here, what you have to do is leverage a different set of technologies, but deliver a capability to the business to predict or to solve complex decision making routines.

The foundation for all of this is data, emphasized O'Brien. "The point here is that we have a conceptual architecture. We have a strategy that says our architecture for data and analytics will focus on delivering analytic capabilities." As you're looking at projects, said O'Brien, you can consider whether you need to get some self-service involved so that you can explore the data and validate whether it is even feasible. You can consider whether the business understands all of their requirements, and then also have a second mode where you build the data models, the data pipelines, and transformations. "And so, conceptually, that's how we can look at our agile projects and say, Which components or capabilities do we  need to have in the architecture available to enable a capability?"

Videos of full presentations from Data Summit Connect Fall 2020, a 3-day series of data management and analytics webinars presented by DBTA and Big Data Quarterly, are also now available for on-demand viewing on the DBTA YouTube channel.