Filters
The Filters section is where you can analyze the process while considering a limited subset of cases that answers to specific user requests. You can set the filters in the interactive panes.


-
Filter Operation
Gives the option to include or to exclude the cases that respond to the filter attributes. You can either:- Match the filter anywhere within a case at the start of the case (only the first activity of each case is considered to decide whether to apply the filter).
- Or at the end of the case (only the last activity of each case is considered to decide whether to apply the filter).
-
Event Attribute
Select the filter’s attribute (for example, activity, resource, role, process-ids, custom field). -
Attribute value
Filters by a specific attribute value. -
Duration
Filters in terms of time, on condition that the Event attribute selected is “Process”.- Over X days
Filters cases that have a duration of more than X days. - Under X days
Filters cases that have a duration of less or exactly X days. You can select the granularity of the time filter (Days/Hours/Minutes/Seconds).
- Over X days
-
Time span
Determines the time period you want to consider for the analysis; the time span filter can be applied on cases or activities. By turning on the Bound time option:- Filter on cases (process as Event Attribute)
The time span filter considers only the instances that started and ended within the selected dates. - Filter on activities
- Flat Process: The time span filter considers only the instances in which the selected activity started and ended within the selected dates.
- Multilevel Process: The time span filter considers only the cases in which all The activities of the selected type started and ended within the selected dates (for example, an invoice with two orders, created in two different time windows).
- Filter on cases (process as Event Attribute)
-
Running or Complete
Focuses on running cases or on completed cases. Running cases are automatically identified by IBM Process Mining. You can fine-tune the algorithm by using the “End activities” in the settings. -
Simulation
Excludes simulated cases or exclude As-Is (uploaded from Datasource) cases. IBM Process Mining can automatically distinguish the simulated scenarios from the imported data sources. -
Conformant or Non-conformant
Includes or excludes all cases that are conformant to the reference model (reference model must be imported to allow this option). -
Process flow pattern
Choose the flow precedence conditions and the rework boundaries, based on a selected attribute
(for example, Activity), by using the following syntax:
Depending on which Event attribute is selected, different flow precedence conditions
become available. You can also select the path to be filtered by clicking
and selecting the activities on the process map (in this case the Event attribute is automatically
set to “Activity”). IBM Process Mining filters all cases in which the selected process flow pattern exists.


- Path time filter
After selecting a Process flow pattern based on Activity attribute, you can filter it in terms of path duration.- Over X days
Filters cases in which the selected path has a duration of more than X days. - Under X days
Filters cases in which the selected path has a duration of less or exactly X days.
- Over X days
Tip: In multilevel processes, in which the path can be covered multiple times. The Path time filter is applied if at least one path matches with the duration specification.
For example, let’s include all the cases that have a path duration over 50 days, from Goods Registered to Invoice Registered. The case is included as at least one of its paths has a duration that exceeds 50 days.
In general, filters on multilevel processes consider the relationship between the different entities involved.
For example, in P2P, if you have an order with a requisition (starting with "Requisition Created") and an order without a requisition (starting with "Order Item Created"), but both the orders are registered in the same invoice, they will be considered part of the same case.

Outlier
Outliers are cases and activities whose performance is different from the average.
With outlier filtering, you can remove the outliers from the process to have more uniform performances or you can keep only on the outliers to analyze them.

Note: Outlier filtering is applied at case level (like every filter): therefore, if a certain case contains only an outlier activity, the whole case is included or excluded by applying the filter.
-
Outlier Configuration
Select one or more conditions to filter the outliers:- Activity Service Time
Filter cases in which at least one activity has an outlier service time. - Activity Waiting Time
Filter cases in which at least one activity has an outlier waiting time. - Case Lead Time
Filter cases with outlier lead time.
- Activity Service Time
-
Filter selection: you can choose to consider as outliers
- Lower
Only the performances that are lower than the average. - Upper
Only the performances that are higher than the average. - Both
Both the performances that are higher and lower than the average.
- Lower
-
Outlier Filtering scroll-bar
Decide the strength of the filter. The stronger the filter, the lower tolerance is used while computing what cases and activities are outliers.
KPI
You can filter and visualize cases violating in line with one or more of the selected KPI rules.

The KPI rules are the following:
-
Activity service time
Include or exclude cases that contain at least one activity whose service time exceeds the activity duration threshold (previously defined in the KPI settings). -
Activity wait time
Include or exclude cases that contain at least one activity whose wait time exceeds the activity wait queue threshold (previously defined in the KPI settings). -
Activity throughput time
Include or exclude cases that contain at least one activity whose throughput time (service + wait) exceeds the activity throughput threshold (previously defined in the KPI settings). -
Case lead time
Include or exclude cases whose lead time exceeds the case duration threshold (previously defined in the KPI settings).
Script filters
The Script Filters section is dedicated to developers who intend on building code-based filters.
For more information, have a look to the script filter section.

Managing filters

Save filters as template
You can save a set of filters as a template to reuse it in the future.


-
Load
Apply the selected filter template (other filters are overridden). -
Remove
Delete the selected template. -
Shared
Share your filter template with other users to enable them to see and use the filter in their snapshot of the project.
Removing filters
By selecting an existing filter and clicking “Remove selected filters”, you can remove it.

If a specific filter is not selected, you can alternatively reset all filters.

Edit existing filters
By selecting an existing filter and clicking
, you can edit it.

Activity groups (alias)
![]()
In addition to filters, you can also set aliases on the activities:

By inserting the same alias to multiple activities, if the 2 activities occur within the same case, the second one will be considered a rework.
Note: Aliases are not reset when you reset filters. The aliases must be removed one by one.