Will Scarbrough on 20 Dec 2018 05:25:58
When dealing with data about people, often there is a need to prevent an ability to tie back data to a specific person. For example, feedback data from employees. If Bob says the company is doing a terrible job it is important to see how many say that in aggregate but could blowback to Bob if not kept anonymous.
Even if individual rows can be hidden from the user, when viewing an aggregate value for a small number of records the user can guess who that is. For example, if grouping by location then knowing all 2 people in Boston answer no means you can guess what Bob in Boston had to say.
This request is to implement an ability to hide the aggregate value when the grouping brings the records being aggregated below X. (Best practice for anonymization is fewer than 5 records.) Currently (Dec 2018) we were not able to implement this in DAX, since the DAX does not update dynamically as report selections are changed.