Investigation of the effects of socio-economic indicators on the urban area by spatial and nonspatial techniques

Abstract

The aim of this study is to determine the spatial spread of all the cities with the use of remote sensing technologies in Turkey and to examine the relationship between this spatial extension and socio-economic indicators with the use of Geographic Information Systems (GIS). In the study, the nighttime images taken in 2010 and the socio-economic indicator data of the provinces were used. For determination of urban areas of provinces, the Otsu method, which is a threshold determination method, was used on grey level images. When the spread of the urban areas in the country was examined, it was seen that the urban area sizes are higher in the provinces in the central and western regions. According to the multiple linear regression (MLR) and geographical weighted regression (GWR) analysis conducted between the urban areas and socio-economic indicators, the highest positive effect on the urban area was found to be the literacy rate, which is an indicator of education level. In this study, the results of MLR and GWR were compared. In the GWR analysis, the provincial values ​​were obtained, and the rate of explanation of urban areas was found to be higher differently from the MLR analysis. For the verification of regression models, the data of 2011 were used. At the end of the verification, the accuracy of the GWR analysis was higher.