Process Mining uses data to discover, monitor, and improve ongoing processes running inside organizations. Even though the relatively novel research discipline has matured theoretically, its practical application was treated with restraint by companies and lacks methodological thoroughness to this day. In order to diminish these shortcomings, this thesis aims to develop a process mining project methodology, which guides organizations through the implementation of process mining projects in practice. This work was developed using the Design Science Research Methodology (DSRM) and is based on both a theoretical and practical foundation. To this end, preexisting theoretical methodologies and fundamental literature were instrumented to build the methodology’s rough framework.
Subsequently, a case study analysis enabled the incorporation of further details about specific tools, techniques, and roles typically employed in practical process mining applications. This way, the methodology obtains a deep level of detail, is geared to practical applicability in several economic sectors, and facilitates the entry into the field for beginners. Ultimately, a significant number of domain experts evaluated the methodology in an online survey format, allowing for further improvement of its design and validity.
Table of Contents
1 Introduction
1.1 Problem Identification
1.2 Objectives of this Work
1.3 Research Structure
2 Theoretical Background
2.1 Business Process Management
2.2 Data Mining
2.3 Process Mining
2.3.1 Types and Perspectives
2.3.2 Tools and Algorithms
3 Research Design
3.1 The Research Design
3.1.1 The Design Science Research Methodology
3.1.2 The adapted DSRM
3.2 The Artifact
3.2.1 Attributes
3.2.2 Elements
4 Requirements of a Methodology
4.1 Best Practices
4.2 Challenges
4.3 User Expectations
5 Development of the Preliminary Design
5.1 Analysis of existing Methodologies
5.1.1 Similarities and Differences
5.1.2 Drawbacks
5.2 Development of the preliminary Methodology
6 Practical Refinement
6.1 Case Study Analysis
6.1.1 Literature Review
6.1.2 Template and Case Study Analysis
6.2 Findings of the Case Studies
6.3 Learnings and Limitations of the Case Studies
7 Evaluation
7.1 Evaluation Criteria and Design
7.2 Content and Assessment
7.3 Results and Adjustment
8 Development of the Final Design
8.1 Process Model
8.2 Stages of the Final Methodology
8.2.1 Project Set-up
8.2.2 Data Extraction
8.2.3 Event Log Improvement
8.2.4 Process Mining
8.2.5 Evaluation and Interpretation
8.2.6 Value Creation
9 Discussion and Conclusion
9.1 Summary of the Approach
9.2 Achievements
9.3 Limitations
9.4 Recommendations for Future Work
9.5 General Conclusion
Objectives & Core Topics
This thesis aims to develop a comprehensive, practically applicable, and easy-to-understand project methodology for process mining. The research question addresses the challenge of creating a structured guide that supports organizations in implementing process mining projects, thereby bridging the gap between theoretical potential and practical execution.
- Development of a structured process mining project methodology.
- Application of Design Science Research (DSRM) framework.
- Analysis of existing methodologies and synthesis of best practices.
- Evaluation of the methodology through expert surveys in clinical and research settings.
- Focus on practical requirements, data extraction, and iterative improvement processes.
Excerpt from the Book
Perform discovery analysis
After providing a high-quality event log by applying several pre-processing activities, the methodology proceeds to its centerpiece, the analysis phase. The latter starts with the discovery analysis that aims to create a process model representing the data from the event log in the best way possible. The discovery analysis is typically conducted by the process mining analyst and should be performed as a first analysis since the process model it creates serves as an integral input for the other types of analysis. It is the fundamental part of a process mining project, which was also demonstrated by the fact that each case study carried out such a discovery analysis. Thereby, four quality dimensions defined by van der Aalst (2016a) need to be considered. However, it is a demanding task to choose the right setting of these criteria, as the four dimensions compete with each other and need to be balanced. Namely, they consist of the dimensions fitness, simplicity, precision, and generalization and are defined more specifically in the following. Figure 20 illustrates the dependency and core tasks of the four criteria.
Summary of Chapters
Introduction: Summarizes the problem of lacking methodological structures in practical process mining and defines the thesis objectives.
Theoretical Background: Provides foundational knowledge on Business Process Management, Data Mining, and key concepts of Process Mining.
Research Design: Outlines the Design Science Research Methodology (DSRM) as the chosen approach to structure the artifact development.
Requirements of a Methodology: Identifies essential requirements derived from literature, including best practices, common challenges, and user expectations.
Development of the Preliminary Design: Compares existing methodologies to distill useful components and creates the initial version of the proposed methodology.
Practical Refinement: Uses case study analysis to refine the methodology with practical insights regarding tools, techniques, and roles.
Evaluation: Describes the methodology’s validation through an online expert survey, measuring ease of use, efficiency, and generality.
Development of the Final Design: Presents the final BPMN-based process model and provides detailed guidelines for each of the six methodology stages.
Discussion and Conclusion: Highlights the achievements of the work, discusses limitations, and suggests recommendations for future research.
Keywords
Process Mining, Methodology Development, Design Science Research, Business Process Management, Data Extraction, Event Log, Project Management, Project Methodology, Process Discovery, Performance Analysis, Conformance Checking, Practical Applicability, Process Improvement.
Frequently Asked Questions
What is the fundamental goal of this thesis?
The primary goal is to develop a robust, practical, and understandable project methodology that guides organizations through the implementation of process mining, thereby addressing the current lack of methodological maturity in the field.
Which central topics are addressed?
The work covers the entire life cycle of a process mining project, ranging from problem identification and data extraction to model discovery, evaluation, and organizational value creation.
What is the core research question?
The main research question is: "What could a project methodology look like that supports organizations in implementing process mining projects in practice?"
Which scientific methodology is used?
This thesis employs the Design Science Research Methodology (DSRM), which provides a structured, iterative framework for creating and evaluating IT artifacts like the proposed methodology.
What is covered in the main development section?
The main section details the six phases of the methodology: Project Set-up, Data Extraction, Event Log Improvement, Process Mining, Evaluation and Interpretation, and Value Creation.
Which are the defining characteristics of this work?
The work is characterized by its focus on practical applicability, its derivation from literature-based best practices, and the objective of making process mining accessible to non-experts through a structured, top-down approach.
Why are iteration loops included in the methodology?
Iteration loops are incorporated because process mining is an exploratory task where initial insights often necessitate returning to earlier stages, such as refining data extraction or re-defining research objectives.
How was the methodology validated?
The artifact was validated through a quantitative online survey conducted with 30 domain experts from both research and industrial sectors, ensuring the methodology's practical relevance and ease of use.
- Citation du texte
- Lukas Braun (Auteur), 2020, Process Mining in Practice. Development of a Project Methodology, Munich, GRIN Verlag, https://www.grin.com/document/1266611
