Process Mining with ProM

Process mining is a new and exciting field which combines business process management with data science.Using process mining techniques you can analyse and visualise business processes based on event data recorded in event logs. Process mining provides a critical, process-centric perspective on data, which is not available with classical data mining or machine learning techniques. For example, with process mining you can: analyse how people use public transportation; verify whether the loan application process in a bank is executed correctly; gain insights in customer journeys on a websites and combine this with interactions with different channels such as phone and e-mail; analyse learning behaviour of students in a MOOC to improve course contents. predict when hardware parts are likely to fail Course build up Week 1 introduces process mining and shows you that event data is everywhere. Furthermore, we explain how you can translate event data that you might have for use in our tool. We also discuss how to start investigating the event log and how to filter the event log based on these insights for further analysis. Week 2: discusses process models and how to evaluate how good a process model is with respect to the data. We also demonstrate several techniques to automatically learn a process model from the data, and discuss the pros and cons of each approach. Week 3: covers additional process mining techniques such as conformance checking, performance analysis and social network analysis. In this week we also show process mining application examples. Week 4: does not introduce new knowledge, but allows you to work on the peer assignment where you are asked to write a real process mining report on real-life event data. After this course you can: work with the free and open source process mining tool ProM process raw event data into an event log for further analysis execute core process mining analysis techniques, and correctly interpret the results gain concrete and actionable process insights from (your own) event data

• Joos Buijs