Developing a GIS-based estimation model using spatial statistical tools for land property valuation

The concept of real estate valuation emerged in order to determine the price of a real estate objectively and impartially in the most appropriate way. Today, the development of technological tools and the inadequacy of classical valuation methods to meet the needs have led to the idea of mass real estate valuation. With this thesis study, it is aimed to develop a mass real estate valuation estimation model that produces reliable and accurate results, supported by Geographic Information Systems (GIS), by using spatial statistical tools in the valuation of land properties. In this context, regression analysis was used as a mass real estate valuation method. Basic valuation problems, such as what are the criteria that affect the value of land-qualified real estate and to what extent these criteria affect the value, were discussed with regression analysis, and the statistical results were examined. As a result of the analyses, the criteria affecting the value were reduced and the predictive model with the highest expressiveness was determined. By improving the prediction model with Geographically Weighted Regression (GWR), a model expression power of 85.47% was achieved. The estimation model was applied to all real estate in the study area and value maps were produced and presented.