14 Case Study: Building a BI Project from Spec
In Chapter 13 we argued that good BI projects begin with a written specification — a Business Requirements Specification (BRS), a data specification with quality criteria, and a validation plan — before anyone touches the data. Specs anchor the team to a shared definition of “done,” surface assumptions early, and shorten the feedback loop with stakeholders. In this chapter we put that discipline to work end-to-end. The HR department of the Brazilian courier company that produced the Absenteeism at Work dataset has asked for an evidence-based analysis of who is at risk of chronic absence and what drives it. We move from a one-line stakeholder request, to a written spec, to an analysis in R that loads the data we cleaned in Chapter 6, to a final set of validated deliverables.
Chapter Goals
Upon concluding this chapter, readers will be able to:
- Translate a stakeholder request into a Business Requirements Specification with measurable acceptance criteria.
- Author a data specification, data dictionary, and quality criteria for the Absenteeism dataset.
- Use an AI assistant to draft, critique, and iterate on a project spec.
- Implement an R analysis whose deliverables map one-to-one to the items in the spec.
- Validate completed work against the spec using executable checks before declaring the project done.