Understanding Reader Takeaways in Thematic Maps Under Varying Text, Detail, and Spatial Autocorrelation

TL;DR

The paper investigates how textual framing interacts with map type to shape reader takeaways from thematic maps. It employs a factorial experimental design across map types (choropleth, isarithmic, hexbin), geographic detail, spatial autocorrelation via $MoranI$, and text semantics, analyzed with generalized linear mixed models and proportional odds models in a dataset of $N=103$ participants. Key contributions include empirical evidence that annotation design and map type jointly influence the source, granularity, and semantic level of takeaways, moderated by text–map alignment and spatial structure. The findings offer practical guidelines for designing thematic geospatial representations in journalism and science, clarifying when to prioritize map-driven versus text-driven communication and how to balance detail with semantic depth.

Abstract

Maps are crucial in conveying geospatial data in diverse contexts such as news and scientific reports. This research, utilizing thematic maps, probes deeper into the underexplored intersection of text framing and map types in influencing map interpretation. In this work, we conducted experiments to evaluate how textual detail and semantic content variations affect the quality of insights derived from map examination. We also explored the influence of explanatory annotations across different map types (e.g., choropleth, hexbin, isarithmic), base map details, and changing levels of spatial autocorrelation in the data. From two online experiments with $N=103$ participants, we found that annotations, their specific attributes, and map type used to present the data significantly shape the quality of takeaways. Notably, we found that the effectiveness of annotations hinges on their contextual integration. These findings offer valuable guidance to the visualization community for crafting impactful thematic geospatial representations.

Figures (17)

• Figure 1: Different types of maps (which are bolded in this caption) with varying semantic levels of text as described by Lundgard and Satyanaryan lundgard2021accessible for classifying text in visualizations. A U.S. State-level choropleth map (A) and a county-level choropleth map (B) using descriptive statistics (Semantic Level 2 or perceiver-independent) text annotations. (C) An isarithmic map with text annotations describing complex and overall trends (Semantic Level 3 or perceiver-dependent). (D) A hexbin map, which is a cartogram variant, containing text annotations with external or background information (Semantic Level 4 or perceiver-dependent).
• Figure 2: Experimental design constructed using JMP® software. The input variables—comprising text-map detail alignment, semantic level, map type, and map detail—were systematically varied to rigorously assess both individual and interaction effects. JMP® was employed to optimize the design of experiments, ensuring comprehensive evaluation of all interaction effects. In Experiment 1, all of the maps are at the state level. In Experiment 2, all maps are choropleth maps, which means that the variable Map Type is constant.
• Figure 3: The stimuli construction pipeline. Data were mapped onto GeoJSON objects and thematic map design principles were applied to generate all stimuli. Map design factors were either varied or kept constant as required by our experimental design (Fig. \ref{['tab:table-studydesign']}). Maps were validated for adherence to best practices in thematic map design. Finally, annotations were added with differing semantic levels and reviewed a final time to produce verified stimuli.
• Figure 4: The 6 datasets used in the study. The Moran's I value was calculated by using the choropleth map. *Dataset 2 is from Tolbert et al. tolbert2015us
• Figure 5: Templates used for L3 annotations.