2024/01/18 by Fischer, Manfred M.
paper · doi:10.57938/db8fdbd4-b17e-4835-a6ee-bb9c80690e88
This article views spatial analysis as a research paradigm that provides a unique set of specialised <br/>techniques and models for a wide range of research questions in which the prime variables of interest <br/>vary significantly over space. The heartland of spatial analysis is concerned with the analysis and <br/>modeling of spatial data. Spatial point patterns and area referenced data represent the most appropriate <br/>perspectives for applications in the social sciences. The researcher analysing and modeling spatial data <br/>tends to be confronted with a series of problems such as the data quality problem, the ecological <br/>fallacy problem, the modifiable areal unit problem, boundary and frame effects, and the spatial <br/>dependence problem. The problem of spatial dependence is at the core of modern spatial analysis and <br/>requires the use of specialised techniques and models in the data analysis. The discussion focuses on <br/>exploratory techniques and model-driven [confirmatory] modes of analysing spatial point patterns and <br/>area data. In closing, prospects are given towards a new style of data-driven spatial analysis <br/>characterized by computational intelligence techniques such as evolutionary computation and neural <br/>network modeling to meet the challenges of huge quantities of spatial data characteristic in remote <br/>sensing, geodemographics and marketing. (author's abstract)