Figure 10.

Application of the Local Indicators of Geocoding Accuracy (LIGA) scatterplot to the bladder cancer case data shown in the bottom of Figure 8. Those observations in the upper right hand quadrant contribute a great deal to the spatial weights and change the most under a given amount of positional error. The user has brush-selected 4 locations on the scatterplot (center) which are shown by large circles in the maps of leverage (left) and perturbability (right). Additional effort may now be allocated to more accurately determine the locations of those four residences. True locations perturbed to protect participant privacy.