Data Impact Analysis in Business Processes

Abstract

Business processes and their outcomes rely on data whose values are changed during process execution. When unexpected changes occur, e.g., due to last minute changes of circumstances, human errors, or corrections of detected errors in data values, this may have consequences for various parts of the process. This challenges the process participants to understand the full impact of the changes and decide on responses or corrective actions. To tackle this challenge, the paper suggests a semi-automated approach for data impact analysis. The approach entails a transformation of business process models to a relational database representation, to which querying is applied, in order to retrieve process elements that are related to a given data change. Specifically, the proposed method receives a data item (an attribute or an object) and information about the current state of process execution (in the form of a trace upon which an unexpected change has occurred). It analyzes the impact of the change in terms of activities, other data items, and gateways that are affected. When evaluating the usefulness of the approach through a case study, it was found that it has the potential to assist experienced process participants, especially when the consequences of the change are extensive, and its locus is in the middle of the process. The approach contributes both to practice with tool-supported guidance on how to handle unexpected data changes, and to research with a set of impact analysis primitives and queries.

A Configuration Taxonomy of Business Process Orientation

“…An eternal mantra of organizations is the continuous improvement of their business processes (Tsoury et al, 2019;Zaby and Wilde, 2018). To measure progress, organizations strive for tools or benchmarks, often resorting to maturity models that assume progress occurs in stages (Röglinger et al, 2012;Tarhan et al, 2016).…”

Quality-Informed Process Mining: A Case for Standardised Data Quality Annotations

“…However, qualityinformed process mining implies that process mining techniques explicitly take log quality and repair information into account. his need is reinforced by recent work on data impact analysis [46], which elaborates on how changes to data can impact process analysis. Similary, Pegoraro et al [34,35] explore how uncertainty information about atributes in an event log, whereby the actual value is unknown but a set of potential values is known, can be used for conformance checking.…”

Integrated Exploration of Data-Intensive Business Processes

“…Further recent contributions confirmed the importance of bridging the gap between data and process design and explored some research directions complementary to the one focusing on conceptual modeling proposed in this paper. They deal with: a formal language for modeling process-and information-related concepts and constraints [45]; the support to analyze the impact of unexpected data changes during process executions [46]; an operational framework, mainly based on BPMN and SQL languages, supporting modeling and verification of business process models enriched with data management capabilities [11]. All these recent contributions share a sound theoretical foundation with our approach.…”