Overview
The BPMN calculator automatically generates Business Process Modeling Notation (BPMN) 2.0 standard diagrams from your event log data using advanced process discovery algorithms. It analyzes your process execution data and produces industry-standard BPMN XML files that can be opened in popular BPMN tools like Camunda Modeler, Signavio, and bpmn.io.
Unlike the interactive Process Map calculator which visualizes your process in mindzieStudio, the BPMN calculator creates formal, executable process models that conform to the BPMN 2.0 specification. This makes them suitable for process automation, formal verification, documentation, and sharing with stakeholders who use standard BPM tools.
Common Uses
- Generate formal BPMN process models for process automation and workflow engine deployment
- Create standardized process documentation that can be edited and shared using industry-standard BPMN tools
- Discover process structures including gateways (XOR, AND, OR) and control flow patterns from event data
- Produce executable process models for BPM systems like Camunda, Flowable, or jBPM
- Compare discovered process models across different time periods, departments, or regions
- Export process models for regulatory compliance documentation and audit trails
- Identify parallel activities and concurrency patterns in manufacturing or approval workflows
Settings
Discovery Algorithm: Select which process discovery algorithm to use for generating the BPMN model. Each algorithm has different strengths:
Split Miner (default): Handles concurrency well, detects all gateway types (XOR, AND, OR), and models loops effectively. Best for complex processes with parallelism and real-world event logs with noise. Uses the Ebsilon and Eta parameters for tuning.
Inductive Miner: Guarantees sound models (no deadlocks) and produces block-structured processes suitable for execution. Best for processes requiring formal verification, BPM system execution, and compliance checking. Does not use Ebsilon or Eta parameters.
Ebsilon: Controls concurrency detection sensitivity in the Split Miner algorithm. This parameter determines how the algorithm distinguishes between sequential and concurrent activities based on edge frequency ratios.
- Lower values (0.05-0.1): Require more balanced edge frequencies to declare concurrency, resulting in more sequential models
- Default value (0.1): Good starting point for balanced process models
- Higher values (0.2-0.3): Allow imbalanced frequencies, detecting more parallelism in the model
Only applies when Discovery Algorithm is set to Split Miner. Recommended range: 0.05 - 0.2 for most processes.
Eta: Controls edge filtering threshold in the Split Miner algorithm. This parameter determines how aggressively the algorithm prunes infrequent process paths by calculating a percentile threshold and removing edges below it.
- Lower values (0.2-0.3): Preserve more process variations but increase model complexity
- Default value (0.4): Filters approximately 40% of less frequent paths
- Higher values (0.6-0.8): Create simpler models by removing more infrequent paths
Only applies when Discovery Algorithm is set to Split Miner. Recommended range: 0.3 - 0.5 for balanced complexity.
Diagram Orientation: Controls whether the BPMN diagram flows horizontally (left to right) or vertically (top to bottom).
- Horizontal (default): Standard BPMN layout with process flowing left to right
- Vertical: Alternative layout with process flowing top to bottom, useful for tall narrow displays
Border Label: Optional text label to include in the BPMN diagram metadata. Use this to identify the process, specify the analysis context, or provide additional information about the discovered model. This label appears in BPMN visualization tools and helps distinguish between multiple process models.
Examples
Example 1: Discovering Purchase Order Approval Process for Automation
Scenario: Your IT team wants to automate the purchase order approval process using a BPM workflow engine. You need a formal, executable BPMN process model that accurately represents the actual approval workflow, not just the documented procedure. The model must be sound (no deadlocks) and ready for deployment in Camunda BPM.
Settings:
- Discovery Algorithm: Inductive Miner
- Ebsilon: 0.1 (not used by Inductive Miner)
- Eta: 0.4 (not used by Inductive Miner)
- Diagram Orientation: Horizontal
- Border Label: Purchase Order Approval - Production Model
Output:
The calculator generates a BPMN 2.0 XML file containing a block-structured process model. When opened in Camunda Modeler, the diagram shows:
- Start event: Process begins when PO is created
- Task: Create Purchase Order
- XOR Gateway (split): Route based on amount threshold
- Path 1 (amount < $5000): Manager Review -> Approve -> End
- Path 2 (amount >= $5000): Manager Review -> Finance Review -> Senior Manager Review -> Approve -> End
- End event: Process completes when PO is approved
The model includes proper BPMN elements like start/end events, tasks, exclusive gateways for decision points, and sequence flows connecting all elements. The XML file can be directly imported into Camunda BPM.
Insights: The Inductive Miner produces a guaranteed-sound process model suitable for workflow automation. The discovered model reveals that your approval process has a clear threshold-based decision point (the XOR gateway splitting on purchase amount), with high-value POs requiring additional review steps. This executable model can be imported into Camunda BPM, enhanced with business rules and human task assignments, and deployed for process automation. The soundness guarantee ensures the automated workflow will not encounter deadlocks or execution errors.
Example 2: Analyzing Manufacturing Process Concurrency
Scenario: Your manufacturing process has several steps that you believe can be executed in parallel, but the documented process shows them as sequential. You want to discover the actual concurrency patterns from production data to optimize the process and reduce cycle time.
Settings:
- Discovery Algorithm: Split Miner
- Ebsilon: 0.25 (elevated to detect parallelism)
- Eta: 0.4 (standard filtering)
- Diagram Orientation: Horizontal
- Border Label: Manufacturing Process - Concurrency Analysis
Output:
The calculator generates a BPMN diagram showing the discovered process structure. When opened in a BPMN viewer, the model displays:
- Sequential start: Receive Order -> Check Inventory -> Allocate Materials
- AND Gateway (split): After materials are allocated, three activities occur in parallel
- Fork 1: Prepare Assembly Line
- Fork 2: Quality Pre-Check Equipment
- Fork 3: Generate Work Orders
- AND Gateway (join): All three parallel activities must complete
- Sequential continuation: Assemble Product -> Quality Inspection -> Package -> Ship
- Self-loop: Quality Inspection can loop back to Assemble Product for rework cases
The AND gateways clearly show which activities the algorithm detected as concurrent based on the event data.
Insights: The Split Miner algorithm with elevated Ebsilon (0.25) successfully detected parallelism in your manufacturing process. The AND gateways reveal that preparation of the assembly line, quality equipment checks, and work order generation actually happen concurrently in practice, even though the documented process shows them sequentially. This discovery indicates an opportunity to formalize this parallelism in your process design, potentially reducing cycle time by ensuring these activities always execute in parallel rather than waiting for sequential completion. The self-loop on Quality Inspection accurately models the rework pattern where failed inspections send products back to assembly.
Example 3: Creating Simplified Process Documentation for Executives
Scenario: You need to create clean, high-level process documentation for executive presentation and regulatory compliance. The detailed event log contains many exceptional cases and variations that would make the diagram too complex. You want a simplified model showing only the main process flows that represent the majority of cases.
Settings:
- Discovery Algorithm: Split Miner
- Ebsilon: 0.1 (standard concurrency detection)
- Eta: 0.7 (aggressive filtering for simplification)
- Diagram Orientation: Horizontal
- Border Label: Invoice Processing - Executive Overview (Top 70% of Paths)
Output:
The calculator generates a simplified BPMN diagram focusing on the most frequent process paths. The model shows:
- Start: Receive Invoice
- Task: Match Invoice to Purchase Order
- XOR Gateway (split): Match result decision
- Path 1 (successful match - 85% of cases): Approve Payment -> Schedule Payment -> End
- Path 2 (match failed - 15% of cases): Request Vendor Correction -> Match Invoice to PO -> Approve Payment -> End
- End: Invoice Processed
The high Eta value (0.7) filtered out less common exceptional paths, leaving only the two main process variants that represent 70% or more of all cases.
Insights: By using aggressive edge filtering (Eta = 0.7), the algorithm removed infrequent exceptional paths and produced a clean, understandable process model suitable for executive presentation. The simplified diagram shows that most invoices (85%) follow a straight-through processing path, while 15% require vendor correction before payment. This high-level view is perfect for regulatory documentation and stakeholder communication without the visual complexity of dozens of edge cases. The BPMN XML can be imported into PowerPoint presentations or process documentation tools used by compliance teams.
Example 4: Comparing Regional Process Variations
Scenario: Your organization has three regional offices (North America, Europe, Asia) that should all follow the same customer onboarding process. You suspect significant variation in how each region executes the process. You want to generate BPMN models for each region to identify structural differences and determine which region follows the most efficient process structure.
Settings:
- Discovery Algorithm: Split Miner
- Ebsilon: 0.1 (consistent across all regions)
- Eta: 0.4 (consistent across all regions)
- Diagram Orientation: Horizontal
- Border Label: Customer Onboarding - [Region Name]
Apply filters to segment your event log by region, then run the BPMN calculator three times (once per region) with identical settings.
Output:
North America BPMN model shows:
- Linear sequential process: Application -> Credit Check -> Document Review -> Approval -> Account Setup
- Simple structure with one main path
- Model file: bpmn_north_america.xml
Europe BPMN model shows:
- More complex structure with XOR gateways
- After Credit Check: XOR split based on credit score
- High score: Direct to Approval
- Low score: Additional Manager Review -> Approval
- Multiple decision points creating branching
- Model file: bpmn_europe.xml
Asia BPMN model shows:
- Highly complex with AND gateways showing parallelism
- After Application: AND split for concurrent activities
- Document Review (parallel)
- Credit Check (parallel)
- Reference Check (parallel)
- AND join before final Approval
- Model file: bpmn_asia.xml
Insights: The three BPMN models reveal significant regional process variation despite supposedly following the same procedure. North America follows a simple sequential process, while Europe has added decision logic based on credit scores, and Asia executes multiple activities in parallel. By comparing the models side-by-side in a BPMN tool, you can identify that Asia's parallel approach likely explains their faster average onboarding time (documented in other analyses). The North America model could be enhanced by adopting Asia's parallelism to reduce cycle time, while Europe's credit score decision logic could be valuable for all regions to adopt. These exported BPMN files provide concrete, visual evidence to support process standardization discussions.
Example 5: Validating Process Redesign Implementation
Scenario: Six months ago, you redesigned the expense reimbursement process to eliminate unnecessary approval steps and enable parallel processing. You want to validate that the new process has been implemented as designed by comparing a BPMN model from current data against the intended BPMN design.
Settings:
- Discovery Algorithm: Inductive Miner (for sound, comparable model)
- Ebsilon: 0.1
- Eta: 0.4
- Diagram Orientation: Horizontal
- Border Label: Expense Reimbursement - Current State
Filter your event log to include only cases from the last three months (post-redesign period).
Output:
The calculator generates a BPMN model from recent event data showing:
- Start: Submit Expense Report
- AND Gateway (split): Parallel processing
- Path 1: Policy Compliance Check
- Path 2: Receipt Validation
- AND Gateway (join): Both checks must complete
- XOR Gateway (split): Based on amount
- Amount < $500: Auto-Approve -> Payment -> End
- Amount >= $500: Manager Review -> Approve -> Payment -> End
- End: Reimbursement Complete
You can now compare this discovered model against your intended redesign specifications.
Insights: The discovered BPMN model confirms that the process redesign has been successfully implemented. The AND gateways show that policy compliance checks and receipt validation now occur in parallel as designed, rather than sequentially as in the old process. The XOR gateway confirms the new automated approval threshold for expenses under $500 is working correctly. By comparing this discovered model with the theoretical redesign BPMN created six months ago, you can validate conformance to the new process design. Any structural differences between the intended and discovered models would indicate implementation gaps or process drift that requires correction.
Output
The BPMN calculator generates a complete BPMN 2.0 XML file named "bpmn.xml" that you can download from the calculator results.
BPMN XML File:
- Standard BPMN 2.0 compliant XML format
- Contains both process logic (tasks, gateways, events, sequence flows) and visual layout (coordinates, shapes, edges)
- Includes proper namespace declarations for compatibility with BPMN tools
- Can be imported into Camunda Modeler, Signavio, bpmn.io, and other BPMN editors
- File encoding: UTF-8
Directly-Follows Graph Table:
- Shows the underlying process structure as a table with From, To, and Count columns
- Displays the frequency of each transition between activities
- Provides raw data that the algorithm used to discover the BPMN model
Diagnostic Information:
- Performance metrics showing execution time for each algorithm step (DFG creation, loop detection, gateway discovery, etc.)
- Warning messages if the algorithm encountered unusual process structures or convergence issues
- Informational messages about algorithm behavior and decisions
Visual Elements in BPMN Tools: When you open the generated XML in a BPMN tool, you will see:
- Tasks: Rectangular boxes representing activities from your event log
- Gateways: Diamond shapes representing decision points (XOR), parallel execution (AND), or inclusive choices (OR)
- Events: Circles representing process start and end points
- Sequence Flows: Arrows connecting elements showing process flow
- Layout: Automatically positioned elements in horizontal or vertical orientation based on your setting
The exported BPMN file can be further edited in BPMN tools to add business rules, assign human tasks, configure service tasks, and deploy to workflow engines for process automation.
This documentation is part of the mindzie Studio process mining platform.