Masterarbeit
Advancing Automated Evaluation of Web User Interfaces
Completion
2026/01
Research Area
Students
Rui Feng
Advisers
Dr.-Ing. Sebastian Heil
Description
In modern web development, the design and quality of Web User Interfaces (WUIs) play a crucial role in shaping user experience, usability, and accessibility. While manual evaluation techniques such as heuristic analysis and usability testing offer valuable insights, they are often time-consuming, subjective, and difficult to scale in fast-paced development environments. Automated WUI evaluation tools such as Aalto Interface Metrics (AIM) and machine-learned model like UIClip assess the design quality and visual relevance of a UI given its screenshot. However, these tools exhibit functional limitations: they support only single-WUI evaluations per session or a single metric at a time, and exhibit a lack of intelligent task scheduling and resource management mechanisms for concurrent tasks.
This thesis aims to advance automated WUI evaluation by developing a scalable, web-based platform that supports multi-WUI and multi-metric assessments executed for several users in parallel. The system needs to support load balancing of evaluation tasks and manage their efficient parallel execution, taking into account different execution characteristics of the metrics and models. The underlying complexity of this parallelized evaluation system needs to be interfaced to its users in a suitable way that makes it easy to operate the system without in-depth technical expertise beyond the WUI evaluations.
The objective of this thesis is the creation of a solution or the combination of existing techniques to solve the problem of improving WUI assessments through a web-based system with usability and performance requirements as described above. This comprises the analysis of the state of the art of UI assessment platforms and load balancing approaches as well as the demonstration of the solution by implementation and a suitable experimental evaluation in a pilot study focusing on performance and usability as outlined above.
