Development of Land Price Model with Geographically Weighted Regression on the Existence of Spatial Planning Zones: A Case Study in the Eastern Bandung City

Abstract

One of the methods used to estimate land prices is the Geographically Weighted Regression (GWR). The GWR method is built based on the dependent and independent variables (land prices) (the spatial proximity between the land object and other facilities). However, this study will develop the independent variable by adding a spatial planning zone to provide the complexity of land price estimation. This study proposes an implementation mechanism by setting each zone type as an independent variable. Based on the spatial planning zones in Eastern Bandung City, there are five spatial planning zones. Thus, 15 variables were used in this GWR model, with ten variables from public facilities and five from spatial planning zones. The variables are categorized into worship, industry, government offices, health, sports/recreation, education, prisons, defense offices, terminals, trade and service zones, industrial zones, and low-residential, medium, and high-residential zones. The results of this study indicate that the implementation of the spatial planning zone variable has a better accuracy rate than the GWR model without involving the spatial planning zone variable. The approach with the proposed mechanism gives better accuracy of 8.6%. Spatial planning zone variable can be a new perspective in making a GWR-based land price estimation model in addition to the physical object variable in the form of public or social facilities, especially to improve the quality of the model formed.