Abstract

Are We in the Right Place? What Our Map Accuracies Aren't Telling Us
Track: Imagery
Authors: Henry Glick

Categorical maps are ubiquitous but the effects of locational error on map accuracy are not well understood. I will present the results of ArcMap-based Python scripting in which Monte Carlo simulations were used to assess changes in map accuracy with increasing locational error. Rapid declines, variability, and collinearity in the results emphasize that (1) small errors can have appreciable effects, (2) one-time accuracy assessment can be unreliable, and (3) kappa provides little utility.