privacy
2 TopicsHide Dynamically datas in a matrix
Hello Everyone, yes i've already posted yesterday about a similar problem but did explain i think i oversimplified my problem and it resulted in helpful people answering beautifully my question but it did not solve my problem To explain my problem i'm working on a Dashboard that displays interesting numbers about education. I'm working with several filters and the data is quite sensible. I want to be able to put what we call in France the "Statistic Secret" to prevent anyone to find specifically people that succeeded/ or not, at their diploma. The usual way is to mask any data that could lead someone to identify 5 people or less. So that is what i want to do and here is the matrix i want to apply it on. As you see it is a matrix with 2 levels, "GFE" and "Title of diploma", i will call them Group and SubGroup. For the two integer(not the %) columns that i will name columnA and columnB that is what i want to do: -If there is one value <5 in column A or B, mask both value (so the value <5 cannot be calculated via the %) -If there is only one subGroup in a Group that has his values masked. Mask another SubGroup values in the Group (so it cannot be calculated via the subtotal of the Group) -if there is only one Group that has his values masked, mask another Group values (So the initial Group values cannot be calculated with the total) and all of his SubGroup values (So the second Group masked cannot be reconstituted) On this example all the datas are in the same table but i have other pages of the Dashboard where there are matrixes that joins multiple tables like diplomas name are on a different table than the values I'm very thankful if someone can solve this problem as it is tearing myself up ! Have a nice day668Views0likes1CommentPreventing hostile self service reporting on privacy sensitive data
The root problem I have is to provide self service reporting on privacy sensitive data and prevent hostile attempts to breech privacy by filtering. The data set has low amounts of data which would make report complately useless if distributions does not show small amounts. My aim is to show small amounts only as part of distribution which is big enough and as reporting is self service, the problem need to be addressed at measure, not at the report. Basically what I need is to hide all data that would have row total or column total at visual less than limit constant in case data is not hidden. And that recardless of axis and filter chosen at report. I would start the idea from something like: my_protected_measure = VAR unprotected_measure = ... VAR column_total = CALCULATE( unprotected_measure, REMOVEFILTERS(current visual row dimension) ) VAR row_total = CALCULATE( unprotected_measure, REMOVEFILTERS(current visual column dimension) ) RETURN IF(AND(row_total >= 5, column_total >= 5), unprotected_measure, BLANK()) How can I refer to current visual row and column dimensions at the measure? If that is not possible, is there some out-of-the-box solution for handling privacy data at self service reporting?Solved463Views0likes1Comment