Abstract
This article extends earlier work on visualizing parameter estimates for spatial analysis. Robust parameters are those that are resistant to outliers, and so are order-based rather than moment-based in their derivation. The principles of geographical weighting of descriptive statistics are reviewed and applied to the development and computation of geographically weighted quantile graphs. The principles are also applied to the development and mapping of robust geographically weighted regression models. The robust statistics and parameters developed in this article are especially useful for visual exploratory data analysis because they provide local information in two contexts: with respect to spatial proximity and with respect to proximity in the quantiles of the variable of interest. The suitability of the results for visualization in maps and quantile plots is demonstrated.
This article was originally published with errors. This version has been amended. Please see Corrigendum (http://dx.doi.org/10.1080/15230406.2013.859338)