Abstract

An Empirical Evaluation of Areal Interpolation for Socio-Economic Data
Track: Census and Statistical Data
Authors: Brian Carson

This paper addresses the accuracy of a novel kriging-based areal interpolation methodology introduced in ArcGIS 10.1. Kriging introduces statistical robustness which is often missing from other areal interpolation methods. However, given the newness of this technique, there is little empirical evidence of how a kriging-based approach performs when applied to typically non-stationary socio-economic data, considering that kriging assumes spatial stationarity. To determine the overall accuracy of the new kriging based methodology, interpolated values were systematically compared to known values for a variety of census based socio-economic variables across a variety of urban areas. Comparisons were performed for both continuous and count data.