Mapping the results of local statistics: Using geographically weighted regression

Area of Science:

• Spatial statistics
• Geographic Information Systems (GIS)

Background:

• Geographically weighted regression (GWR) is increasingly used in social, health, and demographic sciences.
• GWR provides location-specific parameter estimates to analyze spatial nonstationarity.
• Mapping GWR parameter estimates presents a significant challenge for researchers.

Purpose of the Study:

• To introduce a novel and simple mapping technique for GWR parameter estimates.
• To facilitate the exploration and interpretation of spatial nonstationarity in GWR analyses.
• To combine local parameter estimates and local t-values into a single, informative map.

Main Methods:

• Development of a new mapping technique for GWR results.
• Integration of local parameter estimates and local t-values.
• Visual analysis of spatial nonstationarity.

Main Results:

• The proposed mapping technique effectively visualizes spatial nonstationarity.
• The combined map aids in the interpretation of local parameter variations.
• The method enhances the exploratory capabilities of GWR.

Conclusions:

• The presented mapping technique offers a valuable tool for GWR users.
• Improved visualization aids in understanding complex spatial relationships.
• This approach enhances the application of GWR in various scientific fields.