The selection of areas for case study research in socio-economic geography with the application of k-means clustering

Abstract

The grouping techniques which are known in statistics are rarely used by geographers to select a research area. The aim of the paper is to examine the potential use of the k-means clustering (partitioning) method for the selection of spatial units (here: gminas, i.e. the lowest administrative units in Poland) for case studies in socio-economic geography. We explored this topic by solving a practical problem consisting in the optimal designation of gminas for in-depth research on the interaction between nature protection and local and regional development in the Polish Carpathians. Particular attention was devoted to defining an appropriate number of clusters by means of the elbow method as well as the pseudo-F statistic (the Calinski-Harabasz index). The data for the analysis were mostly provided by Statistics Poland and covered the period of 1999–2012. The multi-stage procedure resulted in the selection of the following gminas: Cisna, Lipinki, Ochotnica Dolna, Sękowa, Szczawnica and Zawoja. The example described in the paper demonstrates that the k-means technique, despite its certain deficiencies, may prove useful for creating classifications and typologies leading to the selection of case study sites, as it is relatively time-effective, intuitive and available in opensource software. At the same time, due to the complexity of the socio-economic characteristics of the areas, the application of this method in socio-economic geography may require support in terms of the interpretation of the results through the analysis of additional data sources and expert knowledge.