Visual and statistical assessment of spatial clustering in mapped data

Area of Science:

• Geographic Information Systems (GIS)
• Spatial Analysis
• Data Visualization

Background:

• Health maps are increasingly used to analyze regional health variations.
• Limited research exists on the visual perception of spatial patterns within mapped health data.
• Graphical perception theories suggest map interpretation is complex compared to other graphics.

Purpose of the Study:

• To investigate the visual perception of spatial patterns in mapped data.
• To determine how different data representation methods affect map interpretation.
• To explore the relationship between visual assessments of spatial patterns and statistical measures.

Main Methods:

• An experiment was conducted where observers assessed maps for clustering.
• Various spatial patterns and data representation techniques (shading, plotting symbols) were used.
• Statistical measures of spatial autocorrelation were calculated for comparison.

Main Results:

• Different spatial patterns on maps were visually distinguishable.
• The choice of shading and plotting symbols significantly influenced visual perception.
• A learning effect was observed for complex maps.
• Visual assessments correlated significantly, though imperfectly, with statistical spatial autocorrelation.

Conclusions:

• The visual perception of spatial patterns in maps is influenced by both pattern type and data representation.
• Effective map design is essential for accurate interpretation of spatial health data.
• Further research is needed to refine the understanding of visual perception and statistical measures in spatial analysis.