2023 Volume 59 Pages 76-88
This paper aims to summarize research trends in regional population analysis using spatial statistics, which has been accumulating empirical studies in the field of demography in recent years. The paper provides an overview of spatial statistics and an explanation of spatial autocorrelation and spatial heterogeneity as spatial characteristics. It describes research trends in the field of demography, including （1） a new approach using a spatial econometric model as a methodological innovation in testing the diffusion hypothesis in fertility transition, and （2） a local model that allows the relationship between the dependent and independent variables to differ by region in the regression model.
The diffusion hypothesis in fertility transition was revealed by a series of studies conducted by the "European Fertility Project （1963-1986）". Spatial statistics analysis of diffusion hypothesis reveals that both diffusion and socio-economic effects were detected. However, the diffusion effect was more significant than the socio-economic effect, and the results are consistent with previous studies.
The geographically weighted regression model is one of the nonparametric spatial regression models, which expresses the spatial variation of local regression coefficients by applying spatial weighting. There are examples of analysis in regional fertility analysis, small-area population projections, and regional mortality analysis in Covid-19.
In Japan, spatial data has been remarkably increased in recent years, and with the free software such as QGIS, R, and GeoDa for spatial statistics, the scope of application of spatial statistical methods to regional population analysis is on the way to being expanded.
