地理学“空间分析导论”课程设置研究

空间分析是地理学和其他相关学科的重要研究手段与方法,其重要性日益引起关注。结合1960年代以来空间分析的发展脉络及各时期的典型方法,在对国内外40多所高校、组织与机构的空间分析课程大纲对比研究和教学实践的基础上,提出了“空间分析导论”课程的教学大纲、具体内容和实习软件等问题,最后对空间分析课程设置提出几点建议:① 拟定空间分析课程66学时左右,其中至少12学时左右的上机实习,实习软件可以针对不同的数据类型选用ArcGIS、GeoDa和R。② 在地理学相关专业或其他学科领域的本科高年级开设入门级的空间分析导论课程,之后,根据具体情况在研究生阶段开设比如"地统计学"、"空间模型与建模"等专题课程。③ 为进一步完善国内空间分析课程教材,可考虑选择引进国外相关著作,并筛选、开发相关的教学案例,编写详细的上机实习操作指导。

Geography Curriculum:Introduction to Spatial Analysis

In geography and related disciplines, spatial analysis is an important research tool or method, which has attracted increasing attention in geography and many other subjects, such as ecology, public health, environments, crimes, economics, etc. Based on the concepts, contents of spatial analysis, considering the different development stages and characteristics of spatial analysis since the 1960s, this article studies more than 40 colleges, organizations or universities' syllabus of spatial analysis at home and abroad. Together with the teaching practice of spatial analysis, the authors suggest the "Introduction to Spatial Analysis" course syllabus and the related specific contents, such as the detailed contents, the selection of practice software, class hours, etc. Accordingly, some advice on the teaching of spatial analysis in China are taken: 1) The specific syllabus should include: the introduction (including types and characteristics of spatial/attribute data, spatial autocorrelation, modifiable areal unit problem, edge effects, ecological fallacy), analysis based on maps (including overlay analysis, buffer analysis, Voronoi graph, thiessen polygon, network analysis, optimal path analysis, flow analysis and modeling, location-allocation analysis), geometry analysis (including centrography, distance/area measure, etc.), exploratory spatial data analysis (including visualization, scatter plot, stem-leaf plot, histogram, box plot), spatial point pattern analysis (including completely spatial randomness, quadrat analysis, kernel density estimation, nearest neighbor method, G function, F function, K/L function, Monte Carlo simulation, and some related test methods), area objects and spatial autocorrelation (including spatial weight matrix, global statistics, local statistics, joint count statistics, Moran's I, Geary's C and Getis'G/Gi*, local indicators of spatial association), spatial interpolation and geostatistics (including first law of geography, surface modeling, stochastic/determine modeling, inverse distance weighted interpolation, multinomial interpolation, spline interpolation, regionalized variable, semivariogram, ordinary Kriging, simple Kriging, and other Kriging methods), and new approaches, e.g. neural network (NN), evolutionary computation, cellular automata (CA), agent-based modeling (ABM), etc. in spatial analysis. 2) For the practice software, more than one type of software should be used in the class, in which ArcGIS, and free software GeoDa and R are recommend. Some other software about NN, CA, etc. may also be introduced, if time permitting. And 3) the class hours and practice hours are suggested that all the time should be at least 66 h, of which at least 12 h are practice hours with computer.