Analysts want to focus on building sharp reports and visualizations that answer the organization’s key business questions. Engineers concentrate on orchestrating data pipelines and providing trustworthy semantic models. The Power BI thin report architecture provides a structure that allows multiple authors to collaborate simultaneously while leveraging a centralized governed validated model. This architecture approach has been around a while and is supported in the classic Power BI Service. Switching back and forth from a live-connection to local connection required extra steps.

Fabric has become the premier hosting service for reports. Its UI and version control integration features work well with the thin report design. The Power BI Desktop is still the best tool for authoring reports. Authors can update the model and the report in one tool. However, publishing PBIX reports to the service is not a recommended pattern for enterprise report development and sharing. Also, version control is a must for Fabric deployment pipelines.

This presentation focuses on development solutions that work for analysts and engineers who are not familiar with Git integration. Analysts get the report authoring independence they need while maintaining good data governance. I will present an option for working on a thin report and model locally. Also, I will discuss semantic model compatibility levels and the problems it can cause. I will cover code checkin patterns that allow for smooth report version commits and pipeline deployments.