Kırsallık Göstergeleri Bağlamında Türkiye İllerinin Kümelenmesi ve Devinimi
Kırsallığın beşeri, sosyal, ekonomik ve ekolojik değerler yönünden çeşitliliği yerleşme ekosisteminin sürdürülebilirliği açısından önemlidir. Kırsallığın bu çok bileşenli yapısının; nüfus yoğunluğu, tarım veya doğal kaynaklar gibi tek boyutlu kriterler ile belirlenemeyeceği ve politika üretme konusunda yetersiz kalınacağı/kalındığı konusunda uzlaşı söz konusudur. Kırsallığa ilişkin yazında yer alan gerek kentsel/kırsal tanımının belirsizliği, gerekse tek değişkenli sınıflamaların yarattığı sınırlılıkların tartışılması sonrasında Türkiye ölçeğinde kırsallığın sınıflandırılmasına ilişkin bir yaklaşım öngörülmüştür. Ülkenin bağlamsal gerçekleri ve var olan veri altyapısı, değişkenlerin seçimi ve yöntem konusunda belirleyici olmuştur. Kent ve kırsal bölgelerin bütünleşik olarak yeniden değerlendirilmesini öngören bugünün mekânsal gelişim politikaları açısından yerleşmelerin sosyodemografik, ekonomik ve fiziksel bağlamlar gibi çok yönlü ve çoklu değişkenli süreçler ile ele alınması önemlidir. Çalışma ile Türkiye illeri kırsallığının seçilmiş sosyo-demografik, ekonomik ve fiziksel çevre değişkenleri yardımıyla sınıflandırılması amaçlanmıştır. NUTS-3 düzeyinde yapılan çalışmanın veri seti Türkiye İstatistik Kurumu ve CORİNE arazi örtüsü verilerinden elde edilmiş olup; yöntem olarak hiyerarşik olmayan kümeleme yöntemlerinden K-ortalamalar kullanılmıştır. Üç başlıkta ele alınan çalışmada il-altbölge-bölge düzeyinde yapılan mekânsal değerlendirmeler sonucunda 2006 yılından bugüne ülkede hissedilen doğu-batı arasındaki keskinliğin zaman içerisinde kırıldığı, bölgelerin ya da alt bölgelerin daha heterojen yapıya ulaştıkları görülmektedir.
Clustering and Motion of the Provinces in Turkey in the Context of Rural Indicators
The diversity of rural areas in human, social, economic and ecological values is important for the sustainability of the settlement ecosystem. There is a consensus that the multi-component structure of rural areas cannot be determined by one-dimensional criteria such as population density, agriculture or natural resources and that the past/present policies are insufficient. After the discussion on the limitations of the definitions on the concepts of urban and rural and univariate classifications in the literature, a new approach on the classification of rural areas in Turkey was proposed. The nation’s contextual realities and the current data infrastructure were decisive in the variable and methodology selection. The present study aimed to classify the provincial rural areas in Turkey based on selected socio-demographic, economic and physical environment variables. The study was conducted on NUTS-3 level and the dataset was obtained from Turkey Statistical Institute and CORINE land cover data and K-means clustering, a non-hierarchical clustering method, was used. As a result of the spatial evaluations carried out at the provincial-subregional-regional level in the study, which are discussed under three main headings, it is seen that the sharpness of definition between the east and the west felt in the country has been broken over time.
Rurality is important culturally, socially, politically, economically and especially in the context of the future / sustainability of rural areas. The distinction / relationship between urban and rural is one of the important topics of regional integration in Europe (Öğdül, 2010). Instead of the sharpness / clarity of the distinction between rural and urban areas with the multiple classifications required to produce policies on urban and rural areas that are diversified by developments, the degree of urbanity and rurality has begun to be discussed (Cloke, 1977; Cloke & Edwards, 1986; OECD, 1993; ESPON, 2004; EUROSTAT, 2005). Despite the limited reliability of quantitative criteria, international organizations (such as OECD and EUROSTAT) adopt these criteria as they are particularly useful in the definition of rural areas, especially in comparison between regions or between states. Although there is no single definition that is accepted as urban or rural, it can be said that two of the common features in the European rural areas are low population density and agriculture has an important role in the local economy (Ballas, et al., 2003). Both the uncertainty of the definition of urban and rural, and the limitations of the definitions made with a single variable, directed the researchers to more complex methodologies and to determine / use new variables in classifying regions.
In terms of today’s spatial development policies, which envisage the integrated re-evaluation of urban and rural areas, settlements need to be handled through multi-faceted and multivariate processes such as socio-demographic, economic and physical contexts. The aim of the study was to evaluate with selected socio-demographic, economic and physical variables the time-dependent change of rurality of the Turkey’s provinces. In the first part of the study, the theoretical framework of the rural area, variable selection and clustering analysis in the method section, and in the last part, the evaluation of the time-dependent changes of the clusters that emerged after the analysis were made at the provincial, subregional and regional level.
The variables obtained from TUİK and CORINE land cover data constitute the dataset of this study conducted at NUTS-3 level. It has been compiled on the basis of those years for which TÜİK and CORINE can provide common data in order to compare the rural structure of the provinces according to time period. Thus, the data set of the study consisted of 10 socio-demographic, 14 economic and 12 physical environment variables for 2006, 2012 and 2018.
The K-means method, one of the non-hierarchical clustering methods, was used in the study aiming to compare the time dependent change of the rural area at NUTS 3 level. Cluster analysis is a method used to analyze and organize multivariate or large scientific data (Everitt, 1993). According to Shih, et al. (2010), the purpose of clustering is to divide the data that can show a high degree of similarity into several groups. In order to make comparisons between variables and clusters of different years, a single number of clusters has been determined for each year. Considering 81 provinces, the number of clusters was calculated as six.
As a result of the K-Means cluster analysis, the variables of “population density” in the formation of socio-demographic clusters, “the rate of export within the country” in the formation of economic clusters and “proportion of artificial areas within the province” in the formation of physical clusters were the most effective factors. As a result of the clustering made at the provincial level by defining the rural indicators and the spatial evaluations at the provincial-subregional-regional level, it is observed that the sharpness of definition between the east and the west felt in the country has been broken over time, and the regions or sub-regions have reached a more heterogeneous structure.
