Comparing Mapping Techniques for RepresentingCO2 Concentration : Evaluating the Effectiveness of Spatial Visualization Methods for expert and Non-Expert Interpretation

Sammanfattning

This thesis investigates how different visual mapping techniques can effectively communicate CO2 concentration data to both Expert and non-Expert users. The aim is to assess which methods best support an accurate and efficient interpretation of spatial air quality information. Five visual mappings—Point, Hexbin, Cluster, 3D Model, and Polygon—were implemented using Kepler.gl and evaluated through a mixed-method user study involving nine participants. The study included two tasks: locating the lowest CO2 value and ranking images based on the average concentration. Quantitative metrics such as completion time and accuracy were combined with qualitative feedback gathered through interviews. The results showed that Cluster and Hexbin consistently performed well in terms of speed, accuracy, and user preference. The 3D Model underperformed across both tasks, primarily due to visual clutter and occlusion issues. Naïve users tended to be faster but slightly less accurate, while Experts valued more compact and information-rich visualizations. The combination of color palette and base map was also shown to significantly affect perception, with dark-to-light gradients on light backgrounds being the preferred setup. This study highlights the importance of tailoring visualizations to the target audience and offers design recommendations to create accessible and interpretable spatial representations of environmental data.