Dashboard Visualization for Atmospheric Rivers
Prof. Dr. Tatiana von Landesberger
Computer Sciences – Visualisation and Visual Analytics
Idea
Visualizing meteorological data requires effective and efficient interaction capabilities. Meteorological data have specific characteristics that must be considered. This is especially true for multiple aspects of data measurements on Arctic atmospheric rivers. Using visualization for publication purposes, whether in papers or online, requires high data accessibility and readability, which is achieved through a large number of filters and display options.
Our goal was to develop, implement, and evaluate a novel dashboard visualization of meteorological data, as well as improve the dashboard on atmospheric rivers designed during the “Visual Analysis Lab.”
Result
This project revealed that designing effective dashboards requires robust data processing, interactive visualization techniques, and a coherent, carefully structured layout. As the dashboard solutions were developed and refined, it became clear that layout quality plays a central role in ensuring the dashboard is readable and usable and effectively communicates data. Meanwhile, emerging AI-based design tools began offering new possibilities for automating or accelerating layout creation. AI has the potential to empower non-experts to effortlessly design their own dashboards. The questions are ones of quality and limitations. To examine these possibilities and limitations, as well as their applicability to meteorological data, such as atmospheric rivers, we posed an important research question: To what extent can AI contribute to producing high-quality dashboard layouts, and how does its performance compare to that of human designers who rely on established guidelines?
In our paper "AI or Humans: Who Designs Better Dashboard Layouts? An Initial Study" (Link: "https://visxgenai.github.io/subs-2025/1389/1389-doc.pdf"), which was presented at the 1st VISxGenAI Workshop at IEEE VIS 2025 (Link: "https://visxgenai.github.io/"), we examine precisely this question.