Overview
The Select Variants filter allows you to manually select specific process variants (activity sequences) from your event log. A variant represents a unique path through your business process - the specific sequence of activities that a case follows from start to finish. This filter keeps only the cases that match your selected variants, enabling you to focus your analysis on specific process behaviors or patterns that are relevant to your investigation.
This filter is particularly powerful when you want to analyze a subset of process paths without the complexity of the entire process, or when you've identified specific variants through exploratory analysis and want to examine them in detail.
Common Uses
- Conformance Analysis: Select only the standard process variants to analyze how the "happy path" performs, excluding deviations.
- Deviation Investigation: Choose only non-standard variants to investigate why certain cases followed unusual process paths.
- Performance Comparison: Select specific variants to compare their performance metrics (duration, cost, resource usage).
- Process Improvement: Focus on high-frequency variants that represent the majority of your cases to maximize improvement impact.
- Root Cause Analysis: Isolate problematic variants (e.g., those with rework loops) to understand their root causes.
- Process Documentation: Select representative variants to document different process scenarios for training or compliance purposes.
Settings
Variant Names: Select one or more process variants from the list. Each variant is displayed as a sequence of activities (e.g., "Create Order -> Approve Order -> Ship Order"). The list typically shows frequency information to help you identify the most common or relevant variants. You must select at least one variant for the filter to work.
Note: The variant selection uses exact string matching with the case's variant property. Variants are case-sensitive and must match exactly.
Examples
Example 1: Analyzing Standard Purchase Order Process
Scenario: You want to analyze only purchase orders that followed the standard approval path, excluding expedited orders and those with rejections or rework.
Settings:
- Variant Names: ["Create PO -> Approve PO -> Send to Vendor -> Receive Goods -> Pay Invoice"]
Result: The filter keeps only cases that exactly match this variant sequence, removing all cases that followed different paths (such as those with additional approval steps, rejections, or skipped activities).
Insights: This allows you to establish baseline performance metrics for the standard process, such as average processing time and resource utilization, without the noise of exceptions and deviations.
Example 2: Investigating Rework Cases
Scenario: You've identified several variants that contain rework loops and want to investigate why these cases required rework.
Settings:
- Variant Names:
- "Submit Claim -> Review Claim -> Request Info -> Submit Info -> Review Claim -> Approve Claim"
- "Submit Claim -> Review Claim -> Request Info -> Submit Info -> Review Claim -> Request Info -> Submit Info -> Review Claim -> Approve Claim"
Result: The filter returns only cases that followed these specific rework patterns, where additional information was requested one or more times.
Insights: By isolating these rework cases, you can analyze common attributes (claim type, reviewer, amount) to identify the root causes of rework and develop strategies to reduce it.
Example 3: Comparing Fast-Track vs. Standard Processing
Scenario: Your process has both a standard path and a fast-track path for urgent orders. You want to compare their performance separately.
Settings (for fast-track analysis):
- Variant Names: ["Create Order -> Fast Track Approval -> Immediate Ship -> Express Delivery"]
Result: The filter keeps only cases that used the fast-track process.
Insights: You can now calculate the true cost and duration of fast-track processing separately from standard orders, helping you make informed decisions about when fast-tracking is justified.
Example 4: Focusing on Top 3 Process Variants
Scenario: After running the Variant DNA calculator, you discovered that 3 variants account for 80% of all cases. You want to focus your improvement efforts on these high-volume paths.
Settings:
- Variant Names:
- "A -> B -> C -> D" (45% of cases)
- "A -> B -> E -> D" (25% of cases)
- "A -> C -> D" (10% of cases)
Result: The filter keeps the 80% of cases that follow these three common paths, removing the long tail of infrequent variants.
Insights: This focused view allows you to optimize the processes that affect the majority of your cases, maximizing the return on your improvement initiatives.
Output
The filter returns a new dataset containing only the cases whose process variant matches one of the selected variant names. Each returned case preserves all its original events, attributes, and timestamps.
If no cases match any of the selected variants, the filter returns an empty result set.
Technical Notes
- Filter Type: Case-level filter (removes entire cases, not individual events)
- Variant Matching: Uses exact string matching with the case's variant property
- Selection Logic: Cases are kept if they match ANY of the specified variant names (OR logic)
- Validation: The filter requires at least one variant to be selected and will throw an error if the variant list is empty
- Helper Methods: The filter provides helper methods to retrieve all unique variants and detailed variant information including frequency data
This documentation is part of the mindzieStudio process mining platform.