Back in the day, when we created BI solutions, reports and data models were separate. If you created a cube or Tabular model with Analysis Services, it was developed with Visual Studio and deployed to a server. Reports could be authored and deployed to the report server, separately. Now, with Power BI Desktop, you author your data model and report in the same development space and then deploy the whole kit and kaboodle to the service at once.
Secretly, Power BI actually separates the report from the data model (called a dataset in the service) and gives them both the same name. This is very convenient for self-service projects because it is quick and easy to make changes to the data model, queries and measures if you need to make a report enhancement. This is all well and good for small, one developer projects but what about those larger scale solutions where the data model and reports are developed by different folks, or when multiple reports are connected to the same data model?
At what point does it make sense to separate the data model and reports into separate files?
Read on and let Paul illuminate.