#DCR Process Mining

DCR Path History

Generally DCR Path History displays how people have actually worked and used the system. You might expect all people to work in a similar manner but looking at real data might surprise you. As sketched below you might have an idea of how the work is done, the design, but you might get real evidence from data that provide you with new insight.

DCR Path History is part of the DCR Process Intelligence Suite. It provides insight into how business processes are actually executed.

DCR Path History use a unique algorithm based on DCR Process Models to analyze the various paths and show them in a compressed manner. Instead of trying to draw all the possible paths we record the state of each path and thereby makes it possible to re-use existing states. The result is a more condenced view of all the routes taken.

When designing business processes it is important that laws and rules are followed, that people work compliant according to rules and regulations. However despite strict rules and regulations there can still be many options of how the work is actually done.

Path History makes it possible to analyze a set of flows or process instances. The result is presented online and can be exported by the user.

Now we extend Path History so you can not only analyze flows but also log files from systems such as DCR Open Case Manager or KMD WorkZone. A log file must either follow the DCR XML Log format or the generic XES standard format.

Path History now makes it possible to analyze the data.

When the applicated is launched from the flows tabs we now present the layout outlined below. Try it yourself from this link:

DCR Path History

DCR Path History can be started from the DCR Dashboard for a specific process model.

Simply click the Analyze Flows button. This will open DCR Path History in a new window.

On the right you see three different areas of information:

  1. Data - default we load all non-Forbidden flows as input data for Path History. It is possible to load a log with traces. If you load a log with activities not found in the process model they are added to the model.
  2. Traces - shows the sceanrios ordered by their likelihood. If loading a log of traces the data is analyzed and identical traces grouped together.
  3. Activities - shows activities found in the process model or traces

On the left you show the first selected flows as default. You can add more and more traces to the analysis ending up analyzing all the data.

The most common path is shown with solid colors whereas un-common paths are shown with light colors. The end states are shown in square boxes whereas other transitions are shown with boxes with rounded corners.

We notice that 88% of the traces take the path Fill out an expense report --> Approve --> Pay out.

We notice that 6% af the traces result in the employee withdrawing the expense report. 5% of the times they withdraw them before approval, and 1% after rejected by the manager.

We also notice that 6% of the expense reports get rejected, but later gets approved and paid out.

You can choose to hide activities from the analysis. If we choose to hide the Reject activity we get the following result:

This is useful if we do not really case about the path where an activity is rejected and later approved.

Download DOT file for further analysis

It is possible to download the DOT file for further analysis. In the file menu choose Download DOT. Find information about the DOT format on wikipedia.

We use thra GraphWiz libraries to assist in the visualization.

Visualizing the flow of activities

Path History can also visualize how activities occur over time. In order to do so you need a real process log with timestamps.

In Path History load the log and click on the Annotate button in the bottom of the screen:

Once you do this the data will be analyzed and displayed as sketched below:

DCR Process Intelligence - Timeline annotation

Related Release notes

Release 6.4

  • Welcome to digital on-boarding and exercises
  • Ability to find certified DCR experts
  • Making it easier for modellers - Use drag-n-drop to nest activity

Related Videos

Related Articles

DisCoveR: Accurate & Efficient Discovery of Declarative Process Models