A retail analytics company maintains a centralized Amazon S3 data lake managed by AWS Lake Formation. The company wants to build visual dashboards in Amazon QuickSight by joining data from the S3 data lake with live transactional data stored in an Amazon Aurora MySQL database. The marketing team should only be able to see approved columns from the combined dataset. The company wants to minimize operational complexity while enforcing fine-grained, column-level access control. What is the most efficient way to accomplish this?