From Insights to Interface: Students' Expectations for AI-Assisted Learning Dashboards

The increasing complexity of educational environments and the diverse needs of learners have highlighted a significant gap in effectively utilising data to inform timely and personalized instructional decisions. However, current learning analytics systems often struggle to provide actionable insights that educators can easily interpret and apply, leading to missed opportunities for meaningful student interventions and support. This paper seeks to bridge the existing gap in achieving decisions and actions based on the analytics displayed on learning analytics dashboard by exploring strategies to transform raw educational data into practical, real-time analytics into personalised outcomes. We report on findings from a comprehensive study that used a user-centered approach to develop effective learning analytics dashboards (LADs) for tertiary education. Our research focuses on designing actionable analytics LADs that cater specifically predictive, and prescriptive analytics capabilities, emphasizing personalized user experience in education. Through co-design sessions, we gathered insights from students to understand their needs and requirements when interacting with learning analytics systems focused into actionable analytics. Using this approach, we identified significant stories influencing how students comprehend and utilize visualized information. Based on these insights gained from the stories, we proposed design features tailored to different analytics capabilities, which can guide the design of user interface, and evaluated them through expert reviews and user evaluations. Our study highlights the challenges and opportunities in creating user-friendly LADs and outlines effective methods for designing and implementing these dashboards for diverse audiences in educational settings.