Preserving attribute value differences of neighboring regions in classified choropleth maps

Beyond many other map-reading tasks by using choropleth maps, the visual detection and comparison of attribute value differences of neighboring polygons are of major interest. These tasks are supported not only through an appropriate color scheme, but also through a suitable data classification. However, common classification methods perform a division into intervals only along the number line, with that ignoring the required spatial context. This is our motivation to present a new measure, the Edge Preservation Index. First examinations proved the logical behavior and desired expressiveness of this index, both numerically (in comparison with a different index in use, the Boundary Accuracy Index) and visually (showing a significant increase in edge information).