How to Graph Process Mining Results in Microsoft Power BI

How to Graph Process Mining Results in Microsoft Power BI

Process mining analyzes event logs to map how business processes actually run — not how they were designed to run. Power BI and Microsoft Process Advisor let you import event log data, build process flow visuals, and create bottleneck dashboards. This guide covers the data structure, visuals, and DAX metrics you need.

Key facts

• Microsoft Process Advisor (now in Power Automate) is the native tool for process mining in the Microsoft ecosystem.
• Power BI visualizes process mining output using the Process Mining visual (built-in from March 2023).
• Event logs need three fields: Case ID, Activity Name, and Timestamp. Additional fields (Resource, Cost) enrich analysis.
• Key process metrics: case duration, throughput time, rework rate, and bottleneck activity frequency.
• EPC Group: 29 years of Microsoft consulting, 10,000+ deployments, Microsoft Solutions Partner (core designations).

What is process mining?

Process mining discovers real process flows from system event logs. It shows how a business process — like invoice approval or patient intake — actually executes step by step.

Most organizations find 20–40% more process variants than their documented procedures show. Process mining makes those variants visible. You can then fix rework loops, cut wait times, and reduce compliance risk.

Event log structure: what Power BI needs

Process mining works from event logs. Each row in the log represents one activity in a process instance.

Three columns are required for every event log.

• Case ID — a unique identifier for a process instance. Example: order number, patient encounter ID, invoice number.
• Activity Name — the step performed. Example: "Order Received," "Invoice Approved," "Claim Submitted."
• Timestamp — when the activity occurred. Use ISO 8601 format (YYYY-MM-DD HH:MM:SS).

Optional enrichment columns improve analysis.

• Resource — who performed the activity (user ID or team name).
• Cost — monetary cost of the activity step.
• Outcome — the final result of the case (approved, rejected, escalated).

Data model for Power BI process mining

Build three tables in your Power BI model.

Table	Columns	Purpose
Events (fact)	Case ID, Activity, Timestamp, Resource	One row per activity event
Activities (dimension)	Activity ID, Activity Name, Category	Activity metadata and grouping
Cases (summary)	Case ID, Start Time, End Time, Duration, Variant, Outcome	One row per complete process instance

Join Events to Cases on Case ID. Join Events to Activities on Activity Name. This star schema supports all standard process mining DAX calculations.

Using the Power BI Process Mining visual

Power BI includes a built-in Process Mining visual since March 2023. Here is how to set it up.

1. Load your event log into Power BI Desktop.
2. In the Visualizations pane, select the Process Mining visual (globe icon).
3. Map the fields: Case ID → Case ID field, Activity → Activity Name field, Timestamp → Timestamp field.
4. The visual renders a process flow map showing activity frequency (node size) and sequence edges (connection thickness).
5. Use the Variant filter to isolate specific process paths. Compare the happy path against rework variants.

Key DAX measures for process analytics

Add these calculated measures to your Cases table for KPI cards and charts.

-- Average case duration in hours
Avg Case Duration (hrs) =
AVERAGEX(
    Cases,
    DATEDIFF(Cases[Start Time], Cases[End Time], HOUR)
)

-- Rework rate: cases with a repeated activity
Rework Rate =
DIVIDE(
    COUNTROWS(FILTER(Cases, Cases[HasRework] = TRUE())),
    COUNTROWS(Cases)
)

-- Throughput: cases completed per day
Daily Throughput =
DIVIDE(
    COUNTROWS(FILTER(Cases, Cases[End Time] <> BLANK())),
    DISTINCTCOUNT(Cases[End Date])
)

Process Advisor integration workflow

Microsoft Process Advisor connects directly to Power BI for richer analysis.

1. Upload event logs to Process Advisor — go to flow.microsoft.com → Process Advisor → New process.
2. Let Process Advisor analyze — it generates a process map, variant list, and bottleneck timeline automatically.
3. Export to Power BI — click the Analyze in Power BI button. Process Advisor creates a pre-built Power BI report connected to your process data.
4. Extend in Power BI Desktop — add your own KPI measures, cross-filter with business data (ERP, CRM), and customize the layout.
5. Publish and share — publish to Power BI Service and set a refresh schedule from your source system.

Common process mining use cases

• Order-to-cash — find manual approval steps and duplicate-entry rework driving DSO above target.
• Procure-to-pay — identify invoice holds, three-way match failures, and payment delays.
• Claim processing (insurance) — compare claim handling time across adjusters and identify rework-heavy claim types.
• Patient intake (healthcare) — map actual patient flow vs designed flow and find wait-time bottlenecks.
• IT service management — analyze ticket escalation paths and identify recurring non-compliance with SLAs.

Frequently asked questions

Does Power BI have a native process mining visual?

Yes. Microsoft added the Process Mining visual to Power BI in March 2023. It renders process flow maps, variant analysis, and bottleneck highlighting natively. No AppSource import required.

What is the difference between Process Advisor and the Power BI Process Mining visual?

Process Advisor (in Power Automate) analyzes event logs and generates process maps. It exports results to Power BI for deeper analysis. The Power BI Process Mining visual is the rendering layer — it needs preprocessed data in the correct format to display correctly.

How large can the event log be?

Power BI Process Mining performs well up to 5–10 million events in Import mode. For larger logs, pre-aggregate in Azure Synapse or Fabric before loading into Power BI. Direct Lake mode on Fabric can handle much larger volumes without memory constraints.

Can I connect process mining to SAP or Salesforce data?

Yes. Extract event logs from SAP (using ABAP extractors or SAP Connector for Power BI), Salesforce (via the native Power BI connector), or ServiceNow. Transform the data to the required Case ID / Activity / Timestamp structure in Power Query before loading.

What are variants in process mining?

A variant is a unique sequence of activities for a case. If 80% of orders follow the same five-step path, that is the main variant (the "happy path"). All other sequences are variants. High variant counts indicate process inconsistency and are a primary target for improvement.

Design your process mining dashboard

EPC Group architects design Power BI process mining solutions for Fortune 500 operations teams. Call (888) 381-9725 or request a 30-minute discovery call.