A big challenge is when you have a system like SAP ERP that has many tables with table names that are not related to what they contain. Also, you have thousands of activities for different processes, you'll have to choose the level of granularity you want. In this example you are assuming the process have just the activities you described, but, what if there are others? Domain knowledge is crucial so you don't miss the important ones. Great video btw

Hello, where to get a reliable dataset (publicly available)

Awesome

Could you tell which software required for this project