GIS-based examination of factors affecting residential real estate with geographically weighted regression analysis

Various methods such as classical, statistical and modern are used in real estate valuation activities. Regression analysis is the most preferred technique among statistical methods. This method, which enables both the estimation of the value and the examination of the factors affecting the value, is generally utilized with the classical regression technique. The disadvantage of the method is that it ignores the degree of influence of factors depending on location. For this reason, using the Geographical Weighted Regression (GWR) method, which allows the investigation of the changes in the effects of factors depending on the locations and enables spatial analysis by highlighting the spatial diversity resulting from these changes, will provide much more objective results. In this thesis, the concept of housing valuation has been emphasized. In this context, 372 houses have been identified in 42 neighborhoods in Ortahisar District of Trabzon province in order to examine the change of location-dependent effects of factors affecting housing prices. Housing prices were accepted as the dependent variable, and the location-dependent effect of 21 independent variables on prices was evaluated throughout the study region using the GWR method, and a model that best represents the region was created. The results were mapped using GIS technology and spatial diversity was visualized. It is thought that this study will contribute to both real estate valuation and spatial statistics.