28.08.2025 – Stojko, Laura – Dr. rer. nat. (INF)
„Personalizing User Interfaces of Large Interactive Displays for Intercultural Groups in Semi-Public Areas"
Seite zu Laura Stojko im Computer Science Tree
Seite zu Laura Stojko im Mathematics Genealogy Project
Abstract:
Technology in ubiquitous areas, such as large interactive displays replacing non-digital information boards, addresses diverse user groups with individuals, each with unique design preferences shaped by their cultural background. While personalization focusing on culturally based preferences has been shown to enhance usability and user experience for single-user applications, its implementation in shared user interfaces remains challenging. This thesis investigates how large interactive displays as multi-user applications can be personalized by considering a group’s intercultural design preferences for the user interface design and how this approach influences the user experience.
Therefore, the overarching aim is to develop a personalization approach for large interactive displays in semi-public areas, focusing on intercultural adaptation of key user interface aspects to enhance the user experience. Key challenges include identifying the most suitable group modeling strategy for merging cultural design preferences and investigating ways to adapt key user interface aspects of large interactive displays.
The thesis systematically develops its solution following Peffers et al.’s approved Design Science Research methodology. The developed personalization approach consists of a five-step procedure, ontologies (for group modeling and user interface adaptations), and a system architecture, ensuring a structured implementation. It is demonstrated by implementing it for the CommunityMirrors and evaluated through a mixed method laboratory study involving 24 participants randomly assigned to four groups. The study compared two group modeling strategies for merging cultural user profiles, average without misery and fairness, and investigated the impact on user experience. The participants evaluated three user interfaces during think-aloud sessions, rating each version with the UEQ-S, and compared them during a semi-structured interview, generating over 18 hours of video and audio material. A coding analysis of the interview transcripts revealed interesting findings and improvements.
The study determined the group modeling strategy fairness as the most successful for aggregating cultural user profiles. While the user interface based on the average without misery strategy also outperformed the baseline version, its strict threshold excluded some user preferences, reducing its effectiveness. Overall, the adapted user interfaces, especially those based on the fairness strategy, positively impacted the user experience, with a substantial impact on the pragmatic quality, as shown by the UEQ-S. While the personalization approach aims minimizing active user involvement, the evaluation revealed that user interface aspects with strong interdependencies with others, e.g., “density” in this case, benefit from being customizable to prevent a negative impact on the user experience. The iteratively improved approach systematically incorporates vital variables for the personalization of large interactive displays in semi-public areas by considering a group’s intercultural design preferences.
This work contributes to several research fields within the Human-Computer Interaction community by advancing group-based personalization approaches, benefiting group modeling (group recommender systems), adaptive systems, user interface design, and user experience for ubiquitous technologies. Further, companies interested in deploying large interactive displays in office environments and practitioners focusing on intercultural design challenges may also benefit from these findings.