Process Mining is deprecated with Appian 24.2 and will no longer be available in an upcoming release. Instead, we encourage customers to use Process HQ to explore and analyze business processes and data. |
This page discusses what a filter is, the types of filters available in Process Mining, and when and how to use filters during analysis.
Processes come in all shapes and sizes, as do the event logs that describes how they're operating. This can be a lot of data to work with or try to understand. To help narrow your focus, you can apply filters in Process Mining. A filter restricts the information that's visible on the screen based on your selections. For example, if you wish to see information on activities completed in the last week, this is a filter.
You can use filters to drill down analyses in process mining. For example, it is possible to filter according to certain durations, time spans, or variants to examine these more precisely. When you apply a filter, they affect all displays, statistics, tables, models, diagrams and root cause analyses. Once filters have been applied, the pie chart changes, indicating the percentage of all cases currently visualized and analyzed.
You can apply multiple filters at once, as well as save filter settings for future use.
The Filter dropdown is available in the header bar at the top of the Process Mining window. You can apply filters any time after you've selected a log.
You can hide the filter pane. Click Hide filters to collapse the filter pane. To open the filter pane, click the icon again.
The Number of Variant Groups filter will always appear here by default.
Create a filter to narrow your view in Process Mining. See Types of filters for more information about what you can filter.
To create a filter:
In the header bar, click FILTERS.
After you save a filter, it appears in the filter pane.
Note: You need to add additional information to some filters. For example, if you filter on a categorical attribute, you must provide a specific value to the filter in the filter pane.
Process Mining includes default filters that you can use to show information you're interested in.
Activities filters narrow your view to show cases that contain or do not contain certain activities, endpoints, or relation between activities.
The Include filter lets you select activities that must (or must not) appear in the process flow. A distinction is made between visible and invisible activities. The activities are displayed in two adjacent tables. Use the search and page options to browse the lists.
The Followers filter lets you select process variants that contain or do not contain certain activity sequences. To create this filter:
You can use the follower filter to also determine the time between two activities. Choose whether the path duration should lie within a certain range. To do this, a minimum or maximum duration can be required or allowed. Click Filter to apply your selections.
The Endpoints filter lets you select one or more activities from the visible and invisible activities. In this way, you can filter by variants and activity paths that end (or don't end) with a specific activity.
In the filter, use the cross icon on the activity name to remove this activity as an endpoint.
The Hide filter lets you hide or temporarily remove certain activities from the model and the analysis. This filter contains two tabs: SELECT MANUALLY and SELECT BY FREQUENCY.
On the SELECT MANUALLY tab:
You can also select activities based on their frequency on the SELECT BY FREQUENCY tab:
The Deviation filter lets you drill down on cases that contain or do not contain certain deviations.
Properties filters drill down on cases that contain or do not contain certain attribute values.
The Attribute filter shows you all available attributes. Next to each attribute name, you'll see the attribute classification (numerical, categorical) and attribute type (event and case attribute).
To create an attribute filter:
For each selected attribute, a separate filter is displayed in the filter bar. Categorical attribute values appear in a donut diagram, which displays the distribution of the attributes over the cases. The diagram won't appear if there are more than 10 different values for the attribute. Numeric attribute values appear in a range slider, where you can adjust the minimum and maximum values.
Hover over an area of the diagram to see the value, how many cases have this value, and the percentage of all cases that have this value. Click on a range to set as a filter and see only cases or events that have this attribute.
The 4 eyes principle filter is based on the follower filter. Select the activities in the filter pop-up. The first activity you select specifies the starting point, and the second activity specifies the end point.
Use the radio buttons to select how to display the relationship between the two specific activities:
In order to check whether the principle of dual control has been observed, the test steps should be set as start and end activities. The radio buttons can therefore be used to filter out the process variants that have passed both or none of the test steps.
Click FILTER to apply the configured filter.
The Descriptive Attributes filter lets you filter on cases that contain or do not contain certain attribute values. If your data contains resources, you can check for cases where selected activities have been executed by the same person.
Time filters narrow your view of activities or cases that have been executed, started or ended in a particular period. All time filters include default minimum and maximum time stamps, based on the log you have currently selected.
The Timespan filter lets you view cases occurring during a certain period of time.
To use the filter:
The Case Duration filter lets you focus the analysis on cases whose processing time or case duration is within a defined frame.
To use the filter:
The Time between activities filter lets you drill down on time between two activities.
Use the radio buttons to select how to display the relationship between the two specific activities:
The Activity Duration filter lets you focus the analysis on activities whose duration is within a defined frame.
The Number of Variant Groups filter lets you set which variant groups to analyze. The variant filter always appears in the model views of Process Mining.
The variant filter has a range from 1 to n, where n is the number of identified variant groups. This range appears above the variant slider. Use the two ends of the slider to choose the range of variant groups to display. As the number of variant groups grows, the frequency of cases described by the variants of a group decreases.
Note: Variant groups don't necessarily equate to variants in most cases. Variant groups collect variants according to their frequency. This means that all variants with the same frequency are grouped together. Variant group 1 always represents the most frequent variant(s) and variant group n represents all variants with the least (most likely only one occurrence).
Each filter appears as a card in the filter pane like so:
Depending on the filter you select, there are a few different actions you can take:
Icon | Action |
---|---|
Add additional activities to the filter. | |
Edit the filter properties. | |
Delete the filter. | |
Reset the filter after you've edited it. |
Tip: To delete all filters at once, click REMOVE ALL FILTERS at the bottom of the filter pane.
If you want to use filters in another organization or environment, you can export the filters.
Click Filter settings.
To import filter settings:
You can also find this option in the How would you like to filter? menu, when you are selecting a filter to apply.
You can create a new event log based on the currently filtered data. This is helpful if you want to use the filtered data as a baseline moving forward. The original log file isn't altered. Instead, a new log file is created but only contains the resulting data from the filter currently applied.
To create a new log from the filtered view:
You can share, edit, and delete saved filters from the management hub. The SAVED FILTERS tab displays all of the target models that you've added to Process Mining or that have been shared with you.
You can save filters to make them available for other event logs in your organization or to share with other users.
Your filter settings are saved to use at a later time. Find your saved filters on the SAVED FILTERS tab in the management hub.
You can share filters with the entire organization or with individual users.
To share a filter:
To stop sharing a filter:
On the Filter Management page, you can see the JSON configuration of each filter you have saved.
To view the JSON config for a filter:
In the Used filters window, the filter settings as shown as JSON. You can see the selected filters, values, and ranges.
Filters