Overview
The Same Time Pairs calculator identifies activity pairs that have problematic timestamp data where the temporal order cannot be reliably determined. This specialized data quality calculator analyzes your process data to find activity pairs where events either occur at exactly the same time or where one event has only a date (no time) while another event on the same date has a specific time. These timestamp issues make it impossible to determine which activity truly happened first, potentially affecting process flow analysis and conformance checking.
Common Uses
- Detect data quality issues in event timestamps before performing process analysis
- Identify activities that are frequently logged with identical timestamps
- Find cases where date-only timestamps conflict with date-time timestamps
- Validate data import quality after migrating from legacy systems
- Assess the reliability of temporal ordering in process mining analysis
- Prioritize data cleanup efforts by identifying the most problematic activity pairs
Settings
This calculator requires no configuration settings. It automatically analyzes all activity pairs in your process data to identify those with temporal ordering issues.
Examples
Example 1: Identifying Data Quality Issues in Invoice Processing
Scenario: After importing invoice processing data from a legacy ERP system, you want to verify whether the timestamp data is reliable enough for process mining analysis. Some activities were logged with full timestamps while others only have dates.
Settings:
- No settings required - calculator runs automatically
Output:
The calculator produces a table with the following columns:
- Activity Pair: Shows the activity pair in the format "Activity1 -> Activity2"
- Activity 1: The first activity in the pair
- Activity 2: The second activity in the pair
- Known Order Pair Count: The number of times this activity pair appears in your data where the temporal order CAN be reliably determined (different timestamps with time-of-day values)
The table only shows activity pairs that have timestamp problems. If a pair doesn't appear in the results, it means all instances of that pair have reliable temporal ordering.
Insights:
You discover that "Invoice Received -> Invoice Approved" appears with a Known Order Pair Count of 247. This means there are 247 cases where these activities can be properly ordered, but the calculator identified this pair because there are ALSO cases where:
- Both activities have identical timestamps (logged at exactly the same time)
- One activity has only a date while the other has a date and time on the same day
This tells you that while most instances are fine, there are some cases where you cannot determine whether the invoice was approved before or after it was received, which is a critical data quality issue requiring investigation.
Example 2: Assessing Batch Import Data Quality
Scenario: Your organization performed a bulk data load from a legacy system, and you suspect that many events were assigned the same timestamp during the migration process.
Settings:
- No settings required - calculator runs automatically
Output:
The calculator shows several activity pairs with high Known Order Pair Counts but also flags them as problematic, indicating mixed data quality:
| Activity Pair | Known Order Pair Count |
|---|---|
| Order Created -> Order Validated | 1,523 |
| Order Validated -> Inventory Check | 892 |
| Inventory Check -> Shipping Scheduled | 456 |
Insights:
The presence of these pairs in the output indicates that while thousands of instances have proper temporal ordering, there are also instances with timestamp conflicts. This suggests:
- The bulk import may have assigned default midnight timestamps to some events
- Certain activities might have been batch-processed and logged simultaneously
- Data validation rules were not consistently applied during migration
You should investigate the cases contributing to these problematic pairs to determine whether they represent:
- Legitimate simultaneous execution (rare but possible)
- System clock synchronization issues
- Data migration artifacts requiring correction
Example 3: Validating Real-Time Process Data
Scenario: You are analyzing a manufacturing process where activities are supposed to be logged in real-time. You want to verify that the process control system is correctly timestamping all activities.
Settings:
- No settings required - calculator runs automatically
Output:
The calculator shows only a few activity pairs with very low Known Order Pair Counts:
| Activity Pair | Known Order Pair Count |
|---|---|
| Quality Check -> Package | 3 |
| Package -> Label | 1 |
Insights:
Finding only a small number of problematic pairs with low counts is a positive result. It indicates:
- The vast majority of activity pairs have reliable temporal ordering
- The real-time logging system is working correctly
- Only 4 total instances have timestamp issues (3 + 1)
- These few cases might represent legitimate simultaneous execution or minor system glitches
This gives you confidence that your process data is suitable for detailed temporal analysis, process mining, and conformance checking.
Output
The calculator produces a single table showing only activity pairs that have temporal ordering problems. The table includes:
- Activity Pair column: Displays the directional relationship between two activities (Activity1 -> Activity2)
- Individual activity columns: Shows each activity separately for filtering and analysis
- Known Order Pair Count: Indicates how many times this pair appears with reliable temporal ordering, helping you understand the severity of the issue
The output is interactive - you can click on activity pairs to drill down into the specific cases contributing to the temporal ordering problems.
Important Notes:
- Activity pairs that NEVER have timestamp issues will NOT appear in this output
- Higher Known Order Pair Counts suggest the timestamp problem affects a frequently-occurring activity pair
- An empty result table means all activity pairs in your process have reliable temporal ordering
This documentation is part of the mindzie Studio process mining platform.