Hide data in visualization if too few responses

[maronis]

We have an employee survey and I'm trying to build reports in Power BI. We've promised employees confidentiality so I'd like to suppress data if there are fewer than three respondants.

The output looks something like this for each question:

The underlying data looks like this:

Users have the ability to filter results by timeframe, business unit, etc. I want to avoid a situation where someone filters to a specific time and business unit where there are only one or two survey respondants, making it very easy to identify the specific person's responses. Is there a way for the visualization to be blanked out if slicers/filters result in there being fewer than three respondants? Thanks!

[Ashish_Mathur]

Hi,

In the Visual level filters section of the visualisation pane, apply a filter on the measure of >3.

Hope this helps.

[maronis]

Thanks for the response! So I did think of that - the issue is that because of the way the data is structured, here's what would happen if I applied a filter like that -

Say there was a business unit with four responses, three are 'agree' and one is 'disagree.' If I apply a filter like what you suggest, what would appear in the visualization is a chart that shows three 'agree' responses, but no 'disagree' responses.

Similarly, if there were a business unit with three responses, two of which are 'agree' and one 'disagree' the filter you suggest would return no results at all, when in fact I do want to show results for that business unit because there are three responses in total.

Any other ideas given the clarification above? Thanks again!

[Ashish_Mathur]

Hi,

So then we cannot look at absolute numbers at all - your premise of 3 will not be valid in that case.  We must then compute the numbers of respondent/Total population for that group and filter based on that.

[maronis]

Thanks - is there an easy way to do that? I was trying to think of constructing a measure that would count the total number of respondants and filter based on that but I can't for the life of me think of how I'd construct that.

[Ashish_Mathur]

Hi,

I need your file for that.  Share the download link.

[maronis]

Thanks for the offer to help! I'm afraid I can't share the actual file but we did manage to find a way to restrict the data regardless. Instead of presenting the histogram view, we're just showing an average of each question ('Strongly disagree' = 1, disagree = 2, etc) and by doing that we can restrict the visualization to only show questions that have at least three responses by using a report-level filter.