Chair of Process and Data ScienceCopyright: Wil van der Aalst
The Process and Data Science Group, headed by Prof. Wil van der Aalst, is one of the research units in the Department of Computer Science. The scope of PADS includes all activities where discrete processes are analyzed, reengineered, and/or supported in a data-driven manner. Process-centricity is combined with an array of Data Science techniques. The group’s research and teaching activities can be characterized by the keywords: Data Science, Process Science, Process Mining, Business Process Management, Data Mining, Process Discovery, Conformance Checking, and Simulation.
The group has been established in the context of the Alexander von Humboldt Professorship awarded to Prof. Wil van der Aalst in 2017. The award is Germany’s most prestigious and valuable prize for international researchers. The PADS group supports RWTH’s strategy to further strengthen its Data Science capabilities. The group also closely collaborates with the Fraunhofer Institute for Applied Information Technology (FIT).
Currently, the main research focus is on Process Mining (including process discovery, conformance checking, performance analysis, predictive analytics, operational support, and process improvement). This is combined with neighboring disciplines such as operations research, algorithms, discrete event simulation, business process management, and workflow automation.
Major Research Topics of Interest are
- Foundations of Process Mining. This includes the key capabilities to discover different types of process models from even data. These process models may include temporal, resource, and data perspectives. There should be a continuous interplay between event data and process models to enable conformance checking, performance analysis, and prediction. In the relatively young field of process mining there are still many open problems that require cutting-edge research.
- Dealing with large/distributed/streaming/uncertain Event Data. The abundance of event data enables a range of data science techniques. However, the volume, speed, and variety of data create new challenges. Moreover, data may be uncertain and difficult to extract and handle. Addressing these challenges will help to increase the applicability of process mining in a range of application domains.
- Automated Operational Process Improvement.To develop the technology making it possible to support automated process improvement in a data-driven manner, we need to learn faithful “as-is” models and understand performance characteristics. However, to generate “to-be” models, we also need to inject domain knowledge and elicit goals and constraints. Candidate models needs to be evaluated using operations research techniques (including simulation and queueing analysis).
- Responsible Process Mining. Data are collected on anything, at any time, and any place. Obviously, people are concerned about irresponsible data use. Scientific breakthroughs are needed to ensure fairness, accuracy, confidentiality, and transparency. These challenges also apply to process mining. How to conduct process mining without making unfair conclusions or revealing sensitive information.