Process Mining data model

  • 5 minutes

When you publish a process to Power BI, the system creates a default data source and a corresponding report. This data model comprises various tables, columns, and measures that help you visualize process mining results.

Explore the data model

After connecting Power Automate Process Mining to a Power BI workspace, you can open the prebuilt report in Power BI to explore visualizations and the data model.

Screenshot of the Open in Power BI button.

After you open the report in Power BI, you can download the PBIX file and then open it in the Power BI Desktop app. Alternatively, you can explore the data model directly in Power BI service, which might require the workspace admin to perform more configuration.

The data model consists of the following components:

  • Process data - All process-related data without filtering and calculated measures.

  • View data - Entities that show the process data in the context of the created process analytical view using applied filters, calculated measures, and custom metrics.

  • Visuals data - Entities providing precalculated data that's necessary for the process mining custom visuals to display.

  • Helping entities - Other entities needed by Power BI.

The prebuilt data model, which is centered around process mining data, includes a broad spectrum of metrics. We recommend that you invest time in thoroughly exploring and understanding the existing tables, columns, measures, and relationships before you customize the model or incorporate additional data. The following image depicts the various tables in the process mining data model.

Screenshot of a process mining data model.

Explore the data

After you open the report in Power BI, you can use the Explore this data command to examine the data model, tables, and relationships by creating tabular and matrix representations of the underlying data.

Screenshot of the Explore Process Mining dialog.

This method offers a quick and interactive way to explore the tables and relationships by visualizing the underlying data.

Customize the data model

After gaining a solid understanding of the existing data model, you can begin to customize it. We recommend that you build on the data model that Power Automate Process Mining published by using the Power BI composite data model. You can use the composite data model to include data connections from more than one DirectQuery or import data connection, in any combination that you choose. This approach ensures safe modifications when you need to include more data sources, entities, relationships, or custom DAX.

With the composite data model, you can make customizations atop the existing model. Additionally, you can use the composite data model to:

  • Create new measures and calculated columns.

  • Perform a smooth, regular data refresh.

  • Add data from external sources efficiently.

  • Shield data model extensions from breaking user changes.

New measure example

The prebuilt data model already includes a ReportMeasures table that contains various measures, such as Average Duration or Resource Count. The following new measure reports the average number of events for each case by vendor continent:

Average Events Per Case By Continent =
AVERAGEX(
    GROUPBY(
        Events,
        Events[vendor_continent],
        Events[Case_ID]
    ),
    CALCULATE(COUNTROWS(Events))
)

Then, you can use the Explore this data feature to quickly visualize the results, as shown in the following image.

Screenshot of the Explore this data feature showing a visualization of the results.

Now the new measure is ready to use to customize the report.

Extra data

One advantage of using a custom Power BI workspace is that you can enrich the reports by bringing in external data from other data sources in addition to process mining data. For example, you could bring in the country/region holiday schedule data. Then, you could add new measures to analyze dependencies and correlation between case duration, start and end times, and overlap with the public holidays schedule on each country/region basis.