Prerequisite: None
Mark Peco
CBIP
Analytics Consultant and Instructor
Mark Peco, CBIP, is an experienced consultant, educator, and team builder. With graduate and undergraduate degrees in engineering from the University of Waterloo, he helps clients identify and build the necessary analytics capabilities that will drive business impact.
As a leading practitioner of business intelligence and analytics, Mark is a faculty member of TDWI and teaches companies on a global basis how to implement and govern “intelligent” business solutions. He is CBIP certified at the mastery level and maintains his professional focus at the intersection of business, operations, and technology. Mark has several years of energy industry experience based on a variety of roles in the pipeline, distribution, and production sectors.
Mark lives in the Toronto area and can be contacted via email.
Traditional dimensional modeling techniques are destined to fail in modern BI and analytics programs. The best practices of dimensional modeling were established in the 1990s, before the advent of analytics, self-service, nonrelational data, agile methods, and data governance. The instructor will demonstrate how these expanded interests conflict with classic dimensional approaches and provide you with new techniques for scoping, requirements, design, and refactoring. Nothing is sacred, not even conformed dimensions!
You will also learn how to adapt to new data architectures, changed organizational structures, agile development methods, and nonrelational data management technologies. Both data modelers and non-modelers will learn how to use dimensional concepts to drive requirements that are actionable and useful.
You Will Learn
- How the dimensional model fits—and does not fit—in the expanded scope of modern BI and analytics programs
- The place of dimensional data in expanded data architectures that include self-service, analytics, and data lakes
- Why an enterprise scope can bog down modeling efforts and degrade business value
- When to abandon conformed dimensions, and how to manage expectations accordingly
- Changes to traditional modeling practices, such as establishment of grain, dimension population, and slow change rules
- How business and technologists can work collaboratively for just-enough-modeling and manage efforts so that future refactors don’t get anyone fired
- Simple refactoring techniques and a debt matrix for agile scope management
- Techniques to link dimensional data with big data that lives outside the data warehouse
- What kind of information should be collected and validated so that models are useful and actionable (hint: don’t think “conceptual-logical-physical”)
- A modern requirements process that acknowledges business organizations that often separate data management from BI and analytics teams
- Templates for successful development of dimensional models
Geared To
This course is intended for anyone who contributes to data mart requirements and development, the publication of data sets, or data modeling:
- BI, analytics, and DW program leadership
- Business analysts and requirements analysts
- Data modelers, data architects, and data engineers
- Reporting and data pipeline developers and architects
- Project delivery managers
- “Power users” and business subject matter experts