I had the same confusion at first, but I think I’ve sorted it out. It comes down to poor naming. The “Project” in Data is not the same as the “Project” in Analytics. The way it works is the “Project” in Data equates to a common schema which is applied to two environments. each environment equates to a single Analytics “Project” so that you can ensure you have a consistent plan across environments. With this in place, your instrumenting team will be testing their instrumentation in dev/qa environment based on the same requirements set for the production environment.
One thing that I’ve been thinking about is when you have many projects for say different applications that have some commonalities in schema, should you have one common schema (Data Project) to align all the Analytics projects for these various applications? For example the company has 4 applications, A, B, C, and D and often analyzes the data in cross-project views for various common activities. If a user can like a post in application A, B, and D. and a user can search by genre in application B, C, and D, then should the Data Project have one schema for all 4 applications that includes a “Post liked” event and a “search” event with “filter” property so that the same events and properties are used across all the projects that are expected to have the same nomenclature?