I think you missed measuring exactly what is causing the resource utilization in the database. And instead skipped to creating a version of the database that follows best-practices more closely.
Your view facade approach is reasonable, but whether it improves or degrades performance of the existing application is impossible to predict ahead-of-time. And if it isn't a clear performance win, it's not really a viable approach. There are more targeted ways to improve the analytics queries.
So I think you're back to regular performance analysis, which needs to be based on analysis of query cost backed up by wait stats. Followed by targeted design changes in the database, which may require application changes.
And if you're not already, familiarize yourself with the capabilities of their version of EF to map model columns to database columns with specific data types. Some of the data type changes only require trivial changes on the EF side.
