The Microsoft Graph API library has been around since November 2015 and continues to expand. One of the key endpoints that has centered around utilization reports of the Microsoft 365 ecosystem. These reports can be leveraged to drive better overall adoption of the Microsoft 365 platform and understand trends in use.
Leveraging the Microsoft Graph API, the Microsoft 365 Utilization Metrics app has been the easiest deployed solution to work through this challenge. With a few clicks in AppSource, you can deploy this app within Power BI, update some details on the sematic model, and hit refresh to understand utilization. And while it does show data for the past year, customers frequently want more out of their data.
People forget about the Microsoft Graph API until new technology finds its way into the ecosystem. Enter stage left, Copilot for Microsoft 365, the newest belle at the ball! With new technology coming at a premium price point, companies want to understand utilization to ensure value. The $30 per user per month cost of Copilot suddenly has people wanting to confirm people are leveraging the tool. If not, they want to reclaim the license and get it to someone else.
This gave me the idea to revisit this challenge and see how we could leverage Fabric to make it a reality. While the Microsoft 365 Roadmap has this rolling out soon, a little preparation will allow you to deliver this data when it is finally available.
What are the Microsoft Graph API Reports?
The Microsoft Graph API reports are a collection of endpoints that provide utilization data for users. Aggregated by action, these reports provide visibility to how Microsoft 365 is leveraged. While you can pull data for users in 7, 30, 90, and 180 periods, the data is aggregated by user. This proves a challenge in getting data in a meaningful format if you are looking to track trends over time.
While those periods work, there is another option. Most endpoints allow you to specify an individual date for your query. However, you can only collect that data for the past 30 days. This is a huge restriction and can limit your ability to trend data over a longer period of time.
In the past, if you did not have a data warehousing strategy in place, you were severely limited to what Power BI could do. However, leveraging Microsoft Fabric, this will get you in the right direction pretty quickly. However, with the advent of Microsoft Fabric, we will be able to simplify this process and enable you to track this data with minimal effort!
Preparing Access to Microsoft Graph API Reports
To get ready for the data warehousing process, you must deploy an app registration inside of Entra ID. To get started, sign into Entra ID, navigate to Applications, select App Registration, and start a new registration. Name your new app and keep everything else the same:
Once you click register, copy the Application (Client) ID and Tenant ID from the overview page and save it for later. Next, go to API permissions and click add permission. You will select Microsoft Graph and then application permissions. From there, you will need to add the following permissions:
| Permission Name | Purpose |
| User.Read.All | Provide dimensional data around users within your organization from Entra ID. While User.Read has already been added to the app, it will not work in this situation. |
| Reports.Read.All | Provides access to all reporting endpoints for Microsoft 365. This includes Microsoft Teams, SharePoint, OneDrive, Exchange, and Viva Engage |
Once these permissions are assigned, grant admin consent for your organization. The permissions you added should have green checkmarks next to them:
Lastly, you need to generate a client secret. Go to Certificates & Secrets and add a new Client Secret. Once created, make sure you copy the value for your secret and save it. You cannot get it again and will need to re-create it when you are finished.
Turn off Report Masking
As a security feature, the Microsoft Graph API has user masking available for this data. This helps protect the anonymity of user activity in the tenant. However, if we want to track Copilot for Microsoft 365 utilization, we cannot be anonymous. For some of you, this will not apply. However, this process could require some approvals. Asking early will ensure you can the report mask turned off ahead of time.
To turn off report masking, go to the Microsoft 365 Admin Center and expand settings on the left side of the screen. From there, select org settings and then select reports. A panel will appear on the right side of the screen. Uncheck the option titled display concealed user, group, and site names in all reports:
It is important to note that this feature is either on or off for everyone in the organization for all reports. There are no granular controls for this feature. Therefore, make sure you understand the implications of turning off this setting and communicate it to fellow Microsoft 365 administrators who might be using these reporting features whether by API endpoint or in their respective administrator portal.
What else do I need to know?
This project will require you to understand how to use REST APIs in Dataflows. If you have not read my series on leveraging REST APIs in Power BI, this would be a good place to start. We will need to build a number of connections leveraging this method.
Next, we will have to navigate an authentication process that will require us to use the POST method. I have also written an article on navigating this challenge. This will be a little more complicated, but the code found in this article will be helpful for understanding how we will build a function to drive authentication for these endpoints.
Lasty, you will find yourself writing some M Query code. If you are new to this, the code will be provided to assist you. However, you will want to learn a little more about the language so you understand the method behind the madness.
Conclusion and Next Steps
This setup only takes a few minutes, but it will go a long way in preparing our environment. As mentioned above, if you are not familiar with using REST APIs in Power Query, please take some time to learn about them. I promise you that you will thank me later. This can get very complicated, so understanding the process ahead of time goes a long way!
This will be enough to get thing started for right now. In my next article, we will explore setting up our Fabric environment to query this data and prepare it for the data warehouse.
1 Pingback