Final Thesis: Engineering Workflow Orchestration Interfaces: A Full-Stack Approach to Scalable Pipeline Management, Validated through User-Centered Evaluation

Abstract: As data processing pipelines grow in complexity, effective tools for monitoring and controlling their
execution become increasingly important. Within the MECOIS platform, the Apache Airflow scheduler provides a user interface that is difficult for non-technical users to navigate and requires interaction through a separate
interface rather than integrating into the system directly. This thesis aims to address these limitations by developing a custom user interface within the MECOIS system for orchestrating data-intensive tasks. We present the design and implementation of a user interface based on React and TypeScript to enable direct interaction with the Airflow API and the MECOIS backend. The solution supports triggering workflows, monitoring execution states, accessing logs, and managing task execution through retry and cancellation mechanisms. These functionalities are complemented by a graphical representation of workflow structures as directed acyclic graphs, along with mechanisms for efficient data retrieval, caching, and synchronization. We evaluate usability and effectiveness of the approach through a
qualitative user study. This includes interactive demonstrations and interviews conducted to gather user feedback and identify areas for refinement. The results show improved user interaction and workflow management capabilities. Furthermore, the study revealed shortcomings in the initial implementation and informed subsequent improvements to the user interface.

Keywords: Workflow Scheduling, Airflow, User Study, User Interface Design

PDF: Master Thesis

Reference: Timo Ricardo Meinhof. Engineering Workflow Orchestration Interfaces: A Full-Stack Approach to Scalable Pipeline Management, Validated through User-Centered Evaluation. Master Thesis. Friedrich-Alexander-Universität Erlangen-Nürnberg: 2026.