1. Apply sensitivity labels
In this demo, we will see how to set sensitivity labels on reports and datasets in Power BI. First, we need to navigate to compliance dot microsoft dot com. This is the Microsoft 365 compliance center. If we navigate to Information protection in the Solutions menu and then choose Labels, we can see three labels that I have created: Public, Personal, and Confidential. These are custom labels, meaning you can pick whatever names you choose. After publishing the labels, they become available to Power BI--be warned, however, that it may take up to 24 hours for them to become available.
Now I will navigate to app dot powerbi dot com and access the admin portal from the Settings menu. Here, in Tenant settings, I can navigate to Information protection and allow users to apply sensitivity labels for Power BI content. With this setting enabled, we can use the Microsoft Information Protection sensitivity labels we defined in the compliance center. We can also control who gets access to setting these sensitivity labels, as people with access to set the labels also have access to remove the labels.
Just like publishing labels, this may take up to 24 hours to complete. In practice, it will likely take an hour or two. Now I will navigate to the Wake County workspace. I can see two objects: a dataset and a report. Right now, neither has a sensitivity label set but I can change that. Let’s navigate to the Settings page of the Dataset.
On the Settings page, we can drill into the Sensitivity label section and classify the dataset as one of our options. I’ll choose Personal, indicating that the dataset contains personal data we want to protect. There is a checkbox to apply the label to downstream content and I’ll leave that on as I select Apply.
If I return to the workspace, we can see that the dataset now has a sensitivity label of Personal but so does the Report. That’s because the report uses the dataset – in other words, it is content downstream of the dataset. Let’s suppose I know my report is even more sensitive than the dataset itself. I can navigate to the Settings for the report and control its sensitivity label.
Navigating through the settings pane, I can see that the sensitivity label is currently marked Personal. I’ll change that to Confidential and save the report.
After saving, we can see that the report is marked Confidential and the dataset Personal.
We can also mark datasets and reports in Power BI Desktop. To mark a dataset, we can navigate to Power Query for a table like Restaurants in Wake County.
From here, navigate to Data source settings, choose the data source, edit permissions, and you can select a privacy level.
I’ll set this to Private and then Close and Apply. Importantly, note that this doesn’t change the report’s sensitivity level. To do so, we can change the sensitivity from the Home ribbon menu and choose a setting like Confidential.
Now, when I publish this report, I see that both the report and the dataset have a sensitivity label of Confidential, meaning that the sensitivity label of the report drove the sensitivity label for the dataset. Let’s take what we’ve learned about sensitivity labels and apply it to the following exercises.
2. Let's practice!