Spatial Distribution of Statistical Population Data

This paper points out the essentiality of spatial distribution from statistical population data after analyzing the usual methods for describing urban population density.A framework is proposed for spatial distribution from statistical population data.1) The study region is divided into fine grids;2) Identify a centroid for each zone and assign the population data to this centroid;3) Interpolate a grid surface by using the population data of this set of points and by a interpolation techniques.This paper uses kernel estimation to break down count data published for one fixed set of irregular areal units to a fine grid which can then be re-aggregated to other sets of areas and thus used in conjunction with information on an alternative areal base.Finally,an experiment on a city in Mainland China shows that kernel estimation is feasible and effective for modeling spatial continuous surface of population density for multicentric cities.