Abstract Process mining is an efficient technique that combines data analysis and behavioural process aspects to uncover end-to-end processes from data. Recently, the application of process mining on unstructured data has become popular. Particularly, sensor data from IoT-based systems allow process mining to uncover novel insights that can be used to identify bottlenecks in the process and support decision-making. However, the application of process mining requires bridging challenges. First, (raw) sensor data must be abstracted into discrete events to be useful for process mining. Second, meaningful events must be distilled from the abstracted events, fulfilling the purpose of the analysis. In this paper, a comprehensive literature study is conducted to understand the field of process mining for sensor data. The literature search was guided by three research questions: (1) what are common and underrepresented sensor types for process mining, (2) which aspects of process mining are covered on sensor data, and (3) what are the best practices to improve the understanding, design, and evaluation of process mining on sensor data. A total of 36 related papers were identified, which were then used as a foundation to structure the field of process mining on sensor data and provide recommendations and future research directions. The findings serve as a starting point for designing new techniques, enhancing the dissemination of related approaches, and identifying research gaps in process mining on sensor data.

Protecting personal data about individuals, such as event traces in process mining, is an inherently difficult task since an event trace leaks information about the path in a process model that an individual has triggered. Yet, prior anonymization methods of event traces like k-anonymity or event log sanitization struggled to protect against such leakage, in particular against adversaries with sufficient background knowledge. In this work, we provide a method that tackles the challenge of summarizing sensitive event traces by learning the underlying process tree in a privacy-preserving manner. We prove via the so-called Differential Privacy (DP) property that from the resulting summaries no useful inference can be drawn about any personal data in an event trace. On the technical side, we introduce a differentially private approximation (DPIM) of the Inductive Miner. Experimentally, we compare our DPIM with the Inductive Miner on 14 real-world event traces by evaluating well-known metrics: fitness, precision, simplicity, and generalization. The experiments show that our DPIM not only protects personal data but also generates faithful process trees that exhibit little utility loss above the Inductive Miner.