An Evaluation of Small Area Population Estimation Techniques Using Open Access Ancillary Data

National census data represent the “gold standard” for authoritatively portraying a country's residential population distribution, but their aggregated counts for fixed administrative areas present problems for many geographic information system ( GIS ) analyses. Intelligent areal interpolation algorithms assist by transferring data from one zonal system to another using ancillary data to improve accuracy. All areal interpolation methods make assumptions and generate errors, and performance varies with both specific location and the data inputs used. This study adds to our understanding of the relative merits of alternative methods by comparing dasymetric, street network, and surface‐based models interpolating across two spatial resolutions. It examines the importance of the ancillary data source used to drive the process, particularly the efficacy of open access products. Results from an empirical study show that interpolation accuracy is influenced by the choice of ancillary data input as well as the methodological approach adopted. The strongest overall performance is delivered by dasymetric mapping combined with open access data identifying the locations of buildings. Open access data sets offer considerable potential for widening the use of intelligent population interpolation tools, especially if plug‐in tools to execute these algorithms can be made available for commonly used GIS software packages. 全国人口普查数据代表了一个国家居民人口分布权威描述的黄金标准但以固定行政区域汇总的数据进行GIS分析则存在诸多问题。智能区域插值算法利用辅助数据将数据从一个区域系统转换至另一区域系统以提高数据精度。所有的区域插值方法都作出假设并产生误差且插值性能随着具体位置和数据输入的变化而变化。本研究对密度、街道网络、基于区域模型在两种空间分辨率下插值结果进行比较加强对于可选插值方法优缺点的理解。它检验了应用辅助数据源尤其是开放获取数据源对于驱动这个过程的重要性。实证结果显示插值精度受所选择的输入辅助数据和方法的影响。总体上最好的插值效果来自于采用分区密度图并结合开放获取数据识别建筑位置。尤其是当可以常用GIS软件包中以插件工具执行这些算法时开放数据集可为拓展智能人口插值工具应用领域提供很高的可能。