Overview
The Remove Activities enrichment allows you to permanently remove specific activities from your event log, effectively filtering them out from all cases where they occur. This powerful data cleanup tool physically removes the selected activities and their associated event records from the dataset, reducing the log size and simplifying process analysis by eliminating noise, test data, or irrelevant steps.
Unlike filter-based approaches that temporarily hide data, this enrichment permanently removes the selected activities from the log, recalculating all case statistics and metrics based on the remaining events. This is particularly valuable when you need to focus your analysis on core process steps, eliminate system-generated activities that add no analytical value, or clean up logs containing test activities or data quality issues.
When activities are removed, all events corresponding to those activities are deleted from their cases. If a case consists entirely of removed activities, the case itself remains in the log but will have zero events. The enrichment automatically regenerates the case view after removal to ensure all statistics, paths, and visualizations reflect the cleaned dataset.
Common Uses
- Remove system-generated activities that clutter process views (logging, auditing, notifications)
- Eliminate test activities from production event logs
- Clean up data quality issues by removing activities with incorrect or missing data
- Simplify process models by removing low-value administrative steps
- Focus analysis on core business activities by removing technical or support activities
- Remove automated system activities to analyze only human-performed tasks
- Eliminate deprecated activities that are no longer relevant to current process analysis
- Prepare event logs for specific analyses by removing activities outside the scope of investigation
Settings
Activities To Remove: Select one or more activities from your event log that you want to permanently remove. The dropdown displays all unique activities present in your current dataset, allowing you to select multiple activities for removal in a single operation. Once applied, all events with these activity names will be deleted from the log, and the case view will be regenerated to reflect the changes.
You can select as many activities as needed. Common scenarios include selecting all test-related activities (like "Test Order", "Test Payment"), all system notifications (like "Email Sent", "Log Entry Created"), or specific activities you've identified as irrelevant to your analysis goals.
Important: This operation cannot be undone within the enrichment. If you need to restore removed activities, you must reload your original dataset or remove this enrichment from your enrichment chain. Consider testing the removal with a small selection first, or using filters to preview the impact before permanently removing activities.
Examples
Example 1: Removing System Notifications from Order Processing
Scenario: An e-commerce company's order processing log contains numerous system-generated notification activities (email notifications, SMS alerts, logging entries) that make the process model overly complex and distract from the core business process. The analyst wants to focus only on the actual order fulfillment steps.
Settings:
- Activities To Remove:
- "Email Notification Sent"
- "SMS Alert Sent"
- "System Log Entry Created"
- "Audit Record Generated"
- "Status Change Notification"
Output: All events corresponding to the five selected activities are permanently removed from the log. A case that originally had 15 events (10 business activities + 5 notification activities) now has 10 events. The process model becomes simpler and more readable, showing only the actual business process flow without notification clutter.
Before removal:
- Total Events: 125,450
- Unique Activities: 28
- Average Events per Case: 15.2
After removal:
- Total Events: 83,200
- Unique Activities: 23
- Average Events per Case: 10.1
Insights: By removing notification activities, the company can now clearly see the core order processing flow, making it easier to identify bottlenecks and inefficiencies in the actual fulfillment process. The process discovery algorithms produce cleaner, more interpretable models focused on value-adding activities.
Example 2: Cleaning Test Data from Production Logs
Scenario: A software development team's production deployment includes some test transactions that were accidentally executed in the live environment. These test activities need to be removed to ensure accurate process analytics and reporting.
Settings:
- Activities To Remove:
- "TEST: Create Order"
- "TEST: Process Payment"
- "TEST: Generate Invoice"
- "Test Data Entry"
- "Quality Test Run"
Output: All test-related activities are removed from the log. Cases that were entirely composed of test activities now have zero events but remain in the dataset (they can be subsequently removed using a filter). Mixed cases that contained both production and test activities now show only their legitimate production events.
Insights: The cleaned log now represents only genuine production transactions, ensuring that performance metrics, compliance reports, and process analytics reflect actual business operations rather than being skewed by test data.
Example 3: Simplifying IT Service Management Processes
Scenario: An IT service desk's incident management process includes many automated system activities (auto-assignments, auto-categorizations, auto-escalations) that the team wants to remove to focus analysis on human decision-making and manual interventions.
Settings:
- Activities To Remove:
- "Auto-Assign to Queue"
- "Auto-Categorize Incident"
- "Auto-Calculate Priority"
- "Auto-Update Status"
- "Auto-Send SLA Warning"
- "System Auto-Escalation"
Output: The enrichment removes all automated system activities, leaving only the human-performed activities like "Analyst Review", "Assign to Technician", "Resolve Incident", and "Close Ticket". This reveals the actual human workflow and decision points without the noise of automated system actions.
Insights: By focusing only on human activities, the IT team can better understand where manual effort is required, identify opportunities for further automation, and measure the true time analysts spend on incident resolution rather than including system processing time.
Example 4: Healthcare Patient Journey Analysis
Scenario: A hospital's patient flow analysis includes numerous administrative and billing activities that are not relevant to understanding the clinical care pathway. The quality improvement team wants to analyze only clinical activities.
Settings:
- Activities To Remove:
- "Insurance Verification"
- "Generate Bill"
- "Update Billing System"
- "Send Invoice"
- "Schedule Follow-up Billing"
- "Process Payment"
- "Update Patient Portal"
Output: All financial and administrative activities are removed, leaving only clinical activities such as "Register Patient", "Initial Assessment", "Diagnostic Tests", "Treatment", "Medication Administration", and "Discharge Planning". The resulting process model focuses exclusively on the clinical care pathway.
Insights: The clinical team can now analyze patient care quality, treatment pathways, and clinical decision-making without the complexity of interleaved administrative processes. This enables clearer identification of clinical bottlenecks and opportunities to improve patient care delivery.
Example 5: Manufacturing Process Core Flow Analysis
Scenario: A manufacturing plant's production log contains numerous quality check logging activities and status update activities that are automatically generated. The operations team wants to analyze only the core manufacturing steps.
Settings:
- Activities To Remove:
- "Log Temperature Reading"
- "Log Pressure Reading"
- "Auto-Update WIP Status"
- "Generate QC Report"
- "Update MES System"
- "Timestamp Production Event"
- "Log Operator ID"
Output: The enrichment removes all logging and automated status update activities, leaving only the actual manufacturing operations like "Load Raw Material", "Heat Treatment", "Machining", "Assembly", "Final Inspection", and "Package Product".
Insights: The simplified log makes it easier to understand the actual production flow, calculate accurate cycle times for value-adding activities, and identify where actual production bottlenecks occur versus where data is simply being logged.
Output
The Remove Activities enrichment modifies your event log by permanently deleting all events that match the selected activity names. The impact on your dataset includes:
Removed Events: All events with activity names matching your selection are deleted from their respective cases. These events are completely removed from the dataset and will not appear in any process visualizations, statistics, or analyses performed after this enrichment.
Case Structure Changes: Cases that contained removed activities will have fewer events. The case start time and end time may change if the first or last activity was removed. Case duration is recalculated based on the remaining events.
Updated Statistics: All log-level statistics are recalculated:
- Total event count decreases
- Unique activity count decreases (if removed activities are not duplicated elsewhere)
- Average events per case may change
- Activity frequency distributions are updated
- Process paths and variants are recalculated
Case View Regeneration: The enrichment automatically regenerates the entire case view after removal, ensuring that all derived metrics, process flows, and analytical calculations reflect the cleaned dataset.
Empty Cases: Cases that consisted entirely of removed activities will remain in the log but with zero events. These can be identified and removed using a subsequent filter enrichment if desired.
No New Attributes Created: Unlike other enrichments, this enrichment does not create any new attributes. It modifies the fundamental structure of the log by removing data.
Irreversible Within Chain: Once applied, the removed activities cannot be recovered without removing this enrichment from the enrichment chain or reloading the original dataset. The removal is permanent for all downstream enrichments and analyses.
Impact on Downstream Enrichments: Any enrichments applied after this one will only see the remaining activities. Enrichments that reference removed activities will not find them in the dataset.
The Remove Activities enrichment is a structural transformation of your event log, making it an essential tool for data preparation and cleaning before detailed process analysis. Use it early in your enrichment chain when you need to permanently exclude specific activities from all subsequent analysis.
See Also
- Filter Process Log - Temporarily filter cases or events based on conditions (non-destructive alternative)
- Undesired Activity - Identify and flag unwanted activities without removing them
- Hide Attribute - Remove attributes from view without deleting underlying data
- Allowed Case Start Activities - Ensure cases start with approved activities
- Allowed Case End Activities - Ensure cases end with approved activities
This documentation is part of the mindzie Studio process mining platform.