Overview
The Summarize Values enrichment is a powerful statistical tool that aggregates numeric event-level attribute values across all events within each case, creating a new case attribute containing the sum total. This enrichment is essential for process mining scenarios where you need to understand cumulative values at the case level, such as total order value, accumulated costs, overall quantities, or aggregate scores across process steps.
Unlike simple counting operations that merely tally occurrences, Summarize Values performs mathematical summation on actual numeric data stored in event attributes. This enables comprehensive analysis of financial metrics, resource consumption, performance scores, and any other quantifiable values that need to be totaled across a case's journey through the process. The enrichment intelligently handles both integer and decimal values, preserving data precision based on the source attribute type.
The enrichment also supports advanced filtering capabilities, allowing you to selectively sum values only from specific events that meet your criteria. This targeted aggregation enables sophisticated analyses like calculating total costs only for approved activities, summing quantities only from specific departments, or aggregating scores from particular process phases.
Common Uses
- Calculate total order value by summing line item amounts across all order processing events
- Aggregate total processing time by summing duration values from individual activity executions
- Compute total resource costs by summing labor charges, material costs, or overhead expenses across process steps
- Track cumulative quality scores or defect counts throughout manufacturing processes
- Sum transaction amounts in financial processes to understand total monetary flow per case
- Calculate total quantity produced or consumed across multiple production or inventory events
- Aggregate performance metrics like response times, wait times, or service times across customer interactions
Settings
Filter (Optional): Defines criteria to select which events should be included in the summation. When filters are applied, only events matching all specified conditions will have their values summed. This enables targeted aggregation scenarios such as summing values only from specific activities, time periods, or resource types. If no filter is specified, all events in the case containing non-null values for the selected attribute will be included in the calculation.
New Attribute Name: The name for the new case attribute that will store the calculated sum. This should be a descriptive name that clearly indicates what is being summed, such as "TotalOrderValue", "CumulativeCost", "AggregateQuantity", or "SumOfScores". The attribute name must be unique within the case table and cannot conflict with existing attributes. Choose names that follow your organization's naming conventions and are meaningful to business users.
Attribute Name: The event attribute containing the numeric values to be summed. This must be an existing numeric attribute in the event table (integer or decimal type). The dropdown list shows all available numeric event attributes in your dataset. Common examples include "Amount", "Cost", "Quantity", "Duration", "Score", or any custom numeric attribute specific to your process. The enrichment will automatically skip any events where this attribute has a null value.
Examples
Example 1: Total Order Value in E-Commerce Process
Scenario: An online retailer needs to calculate the total value of each customer order by summing individual line item amounts across all order processing events, from initial placement through fulfillment and delivery.
Settings:
- Filter: Activity equals "Add Item to Order" OR Activity equals "Process Payment" OR Activity equals "Apply Discount"
- New Attribute Name: TotalOrderValue
- Attribute Name: LineItemAmount
Output: The enrichment creates a new case attribute "TotalOrderValue" containing the sum of all LineItemAmount values from the filtered events. For example:
- Case 12345: Events with amounts $49.99, $29.99, $15.00, -$10.00 (discount) result in TotalOrderValue = $84.98
- Case 12346: Events with amounts $199.99, $89.99, $45.00 result in TotalOrderValue = $334.98
The new attribute appears in the case table and can be used for order segmentation, revenue analysis, and customer value assessment.
Insights: This aggregated value enables the business to segment orders by total value, identify high-value transactions requiring special handling, analyze discount impact on order values, and understand revenue distribution patterns across the customer base.
Example 2: Manufacturing Total Defect Count
Scenario: A manufacturing company tracks quality defects at each production station and needs to calculate the total number of defects per production batch to identify problematic batches and analyze quality trends.
Settings:
- Filter: EventType equals "Quality Check" AND DefectsFound greater than 0
- New Attribute Name: TotalDefectsPerBatch
- Attribute Name: DefectsFound
Output: The enrichment sums the DefectsFound values across all quality check events for each batch:
- Batch A1234: Station 1 (2 defects) + Station 3 (1 defect) + Station 5 (3 defects) = TotalDefectsPerBatch: 6
- Batch A1235: Station 2 (1 defect) + Station 4 (1 defect) = TotalDefectsPerBatch: 2
- Batch A1236: No defects found = TotalDefectsPerBatch: 0 (or null if no quality checks)
Insights: Manufacturing teams can now identify batches exceeding quality thresholds, analyze defect patterns across production runs, and correlate total defects with other process variables like shift patterns, equipment maintenance schedules, or raw material suppliers.
Example 3: Healthcare Treatment Cost Aggregation
Scenario: A hospital needs to calculate the total treatment cost for each patient episode by summing individual procedure costs, medication charges, and facility fees across all care delivery events.
Settings:
- Filter: (No filter - include all events with cost data)
- New Attribute Name: TotalEpisodeCost
- Attribute Name: ServiceCost
Output: For each patient episode, the enrichment sums all ServiceCost values:
- Episode P2024-001: Emergency ($500) + X-Ray ($250) + Lab Tests ($180) + Medication ($95) + Consultation ($200) = TotalEpisodeCost: $1,225
- Episode P2024-002: Consultation ($200) + Surgery ($8,500) + Recovery Room ($1,200) + Medication ($450) = TotalEpisodeCost: $10,350
Insights: Healthcare administrators can analyze cost distribution across patient populations, identify cost outliers requiring review, support insurance claim processing, and evaluate the financial impact of different treatment pathways.
Example 4: Procurement Cumulative Savings Calculation
Scenario: A procurement department tracks negotiated savings on each purchase order line item and needs to calculate total savings per procurement case to measure buyer performance and vendor negotiations effectiveness.
Settings:
- Filter: Activity contains "Negotiate" OR Activity contains "Approve Savings"
- New Attribute Name: TotalNegotiatedSavings
- Attribute Name: SavingsAmount
Output: The enrichment aggregates savings across all negotiation and approval events:
- PO-2024-500: Initial Quote Savings ($1,200) + Volume Discount ($800) + Payment Terms Savings ($300) = TotalNegotiatedSavings: $2,300
- PO-2024-501: Competitive Bid Savings ($5,500) + Contract Renewal Discount ($2,000) = TotalNegotiatedSavings: $7,500
Insights: Procurement teams can measure total cost avoidance per purchase order, evaluate buyer effectiveness in negotiations, identify best-performing vendor relationships, and demonstrate the department's contribution to bottom-line savings.
Example 5: Customer Service Interaction Duration
Scenario: A call center wants to calculate the total time spent on each customer case by summing individual interaction durations across all touchpoints including phone calls, chat sessions, and email exchanges.
Settings:
- Filter: Channel in ["Phone", "Chat", "Email"] AND Status equals "Completed"
- New Attribute Name: TotalInteractionMinutes
- Attribute Name: InteractionDuration
Output: For each customer case, the total interaction time is calculated:
- Case CS-8901: Initial Call (15 min) + Follow-up Call (8 min) + Chat Session (12 min) + Email Response (5 min) = TotalInteractionMinutes: 40
- Case CS-8902: Phone Call (25 min) + Escalation Call (18 min) + Resolution Call (10 min) = TotalInteractionMinutes: 53
Insights: Service managers can identify cases requiring excessive support time, analyze the relationship between total interaction duration and customer satisfaction scores, optimize resource allocation based on case complexity, and establish service level benchmarks for different issue types.
Output
The Summarize Values enrichment creates a single new case attribute in your dataset containing the calculated sum for each case. The attribute characteristics depend on the source data type:
Attribute Type: The new attribute will be either integer (Int64) or floating-point (Single) based on the source event attribute's data type. Integer event attributes produce integer sums, while decimal or floating-point attributes produce decimal sums, preserving numerical precision.
Attribute Naming: The new attribute uses the name specified in the "New Attribute Name" setting. This attribute becomes immediately available for use in filters, calculators, dashboards, and other enrichments.
Null Handling: Cases without any qualifying events (based on filters) or cases where all qualifying events have null values in the source attribute will not have a value set in the new attribute. These cases can be identified using "is null" filters if needed.
Integration Points: The generated sum attribute can be utilized across the mindzie platform:
- Filters: Create case filters based on sum thresholds (e.g., TotalOrderValue > 1000)
- Calculators: Use sums in further calculations like averages, ratios, or complex formulas
- Dashboards: Display sum distributions, create value-based segments, or show sum trends
- Process Maps: Color or filter process flows based on cumulative values
- Other Enrichments: Feed sum values into categorization, prediction, or conformance enrichments
Performance Considerations: The enrichment efficiently processes large datasets by calculating sums in a single pass through the event data. For optimal performance with very large logs, consider applying case filters before running the enrichment to limit the scope of calculation.
This documentation is part of the mindzie Studio process mining platform.