본문 바로가기
  • Home

MAUP Analysis of Public Spatial Data: Comparing Administrative, Square, and H3 Hexagonal Grids in Asan

Lee Kyoungsoo 1,  Lee Sangjun 1,  Youngjae Hwang 1

1서강대학교

Accredited

ABSTRACT

This study examines the modifiable areal unit problem by reaggregating population, officially assessed land prices, and points of interest in Asan, Korea, into administrative units, square grids, and H3 hexagonal grids. Methods include descriptive statistics, global Moran’s I, local indicators of spatial association, ordinary least squares, spatial regression, and river-road overlays. Moran’s I ranges from 0.3067 to 0.5268 for population, 0.4963 to 0.7993 for land prices, and 0.2748 to 0.6790 for points of interest. Ordinary least squares fit peaks at the 1㎢ grid (R²=0.728) and H3 level 8 (R²=0.753), but residuals remain spatially autocorrelated throughout. Akaike information criterion comparisons favor spatial lag models at finer resolutions and spatial error models at medium and coarse resolutions. Spatial support changes distributions, clusters, conditional relationships, and spatial-dependence modeling. No grid shape is consistently superior. Spatial-data systems should provide fine-grained data and aggregation metadata, while policy analyses should test robustness across spatial units.

Journal Copyright Policy

No CCL information provided

Citation status

* References for papers published after 2025 are currently being built.