The Event Dictionary Workshop

This workshop is maybe one of the most important at this stage as the purpose is to determine the events we’ll analyze further. So, it must not be under-estimated at it will help in making more sense in the way we map the business need with the data itself. Its purpose is to first check the Step names listed in the previous stage (when doing a Frequency distribution of the Step Names), but also to review them with the business Analyst (or Process Analyst) to:

• Remove the events which are irrelevant or just not necessary. For example we can decide to remove the non-values or non-necessary steps here. In the example below the business analyst has decided to remove the “Order Hold” from the Process flow analysis:

• Change the event label to have something more business oriented (typically when having as Step Names a technical code like below):

• Merging similar steps/events (when the same Step Name comes from different data sources but have different names or labels):

In the example above both steps names will be renamed as “Check Product Stock” (this step Name will have now 12+14=28 occurrences)

• Assigning a label for the unnamed steps. in the example below we can see with the business user which name to provide to the empty step (4 occurrences), for example we can decide it’s a Process error we must monitor like this :

The most efficient way to proceed is to first build a list of all the different and existing Step Names. This list can easily be set up after calculating the Frequency distribution of this SN-KEY field.

Example:

Original Step Name

Stock Check

Customer Check

Checking Stock

Order Complete

Order Hold

Good shipped

Good Accepted

Then we can just duplicate the column :

Original Step Name	Final Step Name
Stock Check	Stock Check
Customer Check	Customer Check
Checking Stock	Checking Stock
Order Complete	Order Complete
Order Hold	Order Hold
Good shipped	Good shipped
Good Accepted	Good Accepted

Now it’s time to call the Business Analyst and review the values of the second column (Final Step Name). This column will contain all the desirable values we want to see in the Process Flow, so as we saw previously we can now change the initial label or remove the step name (by changing the value with a NULL).

Original Step Name	Final Step Name
Stock Check	Product Stock Check
Customer Check	Customer Check
Checking Stock	Product Stock Check
Order Complete	Order Complete
Order Hold	[NULL]
Good shipped	Good shipped
Good Accepted	Good Accepted

This table can now be used as a “lookup table” in the ETL flow which loads the data into the Process Mining solution.

In the ETL flow below we use Alteryx to:

1. Create the initial Step Name Dictionary (first list)
2. Load the data into the Process Mining solution directly (here ABBYY Timeline) by doing a lookup with the altered Step Name Dictionary

The big advantage of this approach is: 

• The re-usability of the Step Name Dictionary in any kind of ETL flow
• The capability to change at any moment the new Step Name values. We can also imagine this table can be directly managed by the business users through a web application for example.
• The scalability of the Process Mining loads