The polygon overlay problem in electoral geography
Résumé
We developed an algorithm for reducing geometric differences between source and target dataset. The algorithm tackles the polygon overlay problem in electoral geography before using areal in- terpolation methods. Our results show that improvement in matching between statistical areas and polling areas can reduce up to half of areal interpolation errors, therefore improving the robustness of analysis of electoral behaviour in their spatial context. This is applied to two case studies: the city of Avignon, France, and the city of Hobart, Australia.