社会科学开始借助于地理信息系统(GIS) 的空间统计分析方法, 来研究社会现象的空间 模式和非常态分布, 研究社会科学中的空间结构问题。人口是城市空间结构演变中最为活跃的因 素之一, 同时也较易获取、容易做到时空连续, 因此往往是衡量城市内部空间结构的重要指标之 一。借助于GIS 技术, 以20 世纪90 年代以来深圳市城市内部人口空间结构研究为例。研究结果 表明深圳市城市内部人口空间结构总体表现为集聚。具体来说: 1990 年和2000 年深圳市人口密 度均以特区内东门街道为最高, 特区外各街道人口密度增长较快; 1990~2000 年人口重心向西北 移动约4km; 1990 年人口密度函数和2000 年人口密度函数均以反函数模型拟合较好, 从市中心 随着距离增加, 人口密度递减; 空间相关性分析结果表明在市中心附近具有较大的向心效应, 带 来周边邻近区域人口的增长, 而在东部郊区人口分布相对稀疏, 出现部分离心效应。

The population distribution reflects the mode of the urban internal population and social spatial structure. Because the population data is very easy to obtain and can be compared in temporal and spatial series. A lot of researchers use population data to reflect the urban internal spatial structure. GIS technologies have been used in spatial structure research in the social sciences frequently recently. In this paper, Shenzhen city is taken as an example in this paper to show how to research urban internal population spatial structure based on GIS. The data of Fifth (in 2000) and Fourth (in 1990) censuses of Shenzhen is used in our research in two aspects. Between 1990 and 2000 every street's population density changed a lot, but the highest value street did not changed. The population center of gravity moved to north west about 4km from year 1990 to 2000. Several population density models were simulated for the population distribution. The method of curve estimation is used in finding the population density mode. The optimal population density mode was the Inverse distribution functions, which means that the further from the center and the spars the city would be. And the exponential model (Clark model) is not the best. The results of spatial auto correlation means that the population in the suburban was growing fast but centripetence in the city center still exists.