A metadata-driven self-service analytics case study

Consider the following scenario for the fictional chain store we used for the guided analysis example:

A business unit has requested a KPI dashboard, but it's not sure exactly what data or KPIs it will need on an ongoing basis and wants to develop and control the dashboard, once it is built, with as little involvement from IT as possible.

We start with the following basic wireframe:

Tip

It's a good idea to set out your analysis design using wireframes, which are simply graphical outlines of the high-level content of your analysis: pages, visualizations, user inputs, and so on. Wireframes help you conceptualize the design and share it with other stakeholders, without any need for actual data or working visualizations.

We need IT to create this wireframe on top of a metadata definition that allows the business unit to define the KPI classes, source tables, and KPIs as required. The detail levels will always be Region, Store, and Department. What...