Overview
The Single Activity Cases filter removes cases from your process log that contain only one event. This filter is designed to help you focus on complete process flows by excluding incomplete cases, system errors, or data quality issues that appear as single-event cases. By filtering out these single-step instances, you can analyze multi-step workflows more effectively and get more accurate insights into your actual process execution.
Common Uses
- Remove incomplete case instances before analyzing process variants
- Clean data by excluding cases that may represent system errors or failed transactions
- Focus analysis on cases that represent actual multi-step process flows
- Improve process mining accuracy by removing noise from single-event entries
- Prepare data for workflow optimization by excluding isolated activities
- Enhance variant analysis by ensuring all cases have meaningful process sequences
Settings
This filter has no configurable settings. It automatically removes all cases that contain exactly one event, keeping only cases with two or more events.
Examples
Example 1: Removing Incomplete Order Cases
Scenario: Your order fulfillment process should have multiple steps (Order Received, Payment Processed, Items Picked, Shipped, Delivered). However, some cases only show "Order Received" with no follow-up events, indicating incomplete or abandoned orders.
Settings:
- No settings required - filter automatically removes single-event cases
Result:
All cases with only one event are removed from the dataset. For example, if you had 1,000 cases total and 150 cases contained only "Order Received" with no other events, those 150 cases would be filtered out, leaving 850 multi-step cases for analysis.
Insights: This filter helps you focus on orders that progressed through your workflow, excluding abandoned or incomplete transactions. This gives you a clearer picture of your actual fulfillment process performance without noise from incomplete cases.
Example 2: Cleaning Quality Check Data
Scenario: Your manufacturing process includes multiple quality checkpoints. Some cases show only a single quality check event, possibly representing spot checks or data entry errors, while complete manufacturing cases show the full production sequence.
Settings:
- No settings required - filter automatically removes single-event cases
Result:
Cases containing only one quality check event are excluded. If your dataset had 500 cases with 75 single-event quality checks, those 75 cases are removed, leaving 425 complete manufacturing cases that went through the full production workflow.
Insights: By removing single-event cases, you can analyze the complete manufacturing workflow without noise from isolated quality checks. This provides more accurate cycle times, bottleneck analysis, and variant frequencies.
Example 3: Preparing Customer Journey Analysis
Scenario: Your customer journey data includes multiple touchpoints (Website Visit, Product View, Add to Cart, Checkout, Purchase). Some sessions only have "Website Visit" with no further interaction, while complete journeys show multiple steps.
Settings:
- No settings required - filter automatically removes single-event cases
Result:
All single-event sessions (bounce visits) are removed from your analysis. If you had 10,000 customer sessions with 4,500 single-event bounces, those are filtered out, leaving 5,500 multi-touchpoint journeys for analysis.
Insights: This filter lets you focus on engaged customer journeys that involved multiple interactions, excluding bounces and single-page visits. This provides better insights into how customers actually navigate through your conversion funnel.
Output
This filter removes entire cases from your dataset. After applying the filter:
- All cases with exactly one event are excluded from the results
- All cases with two or more events remain in the dataset
- The total case count decreases by the number of single-event cases
- Event sequences and timestamps remain unchanged for retained cases
- All case and event attributes are preserved for multi-event cases
The filtered dataset contains only cases that represent actual process flows with multiple steps, making it ideal for process variant analysis, bottleneck detection, and workflow optimization.
This documentation is part of the mindzie Studio process mining platform.