Abstract
In this article, geovisualization is used for the presentation and interpretation of spatial analysis results concerning several house attributes. For that purpose, point data for houses in the region of Attica, Greece are analyzed. The data concern houses for sale and comprise structural characteristics, such as size, age and floor, as well as locational attributes. Geovisualization of house characteristics is performed employing spatial interpolation techniques, kriging techniques, in particular. Spatial autocorrelation in the data is examined through the calculation of the Moran’s I coefficient, while spatial clusters of houses with similar characteristics are identified using the Getis-Ord Gi* local spatial autocorrelation coefficient. Finally, a model is developed in order to predict house prices according to several structural and locational characteristics. In that respect, a classic hedonic pricing model is constructed, which is consequently developed as a geographically weighted regression (GWR) model in a GIS environment. The results of this model indicate that two characteristics, i.e., size and age, account for most of the variability in house prices in the study region. Since GWR is a local model producing different regression parameters for each observation, it is possible to obtain the spatial distribution of the regression parameters, which indicate the significance of the house characteristics for price determination in different locations in the study area.
Highlights
The purpose of this article is to present spatial analysis results and suggest how geovisualization can contribute to their interpretation
Since geographically weighted regression (GWR) is a local model producing different regression parameters for each observation, it is possible to obtain the spatial distribution of the regression parameters, which indicate the significance of the house characteristics for price determination in different locations in the study area
A spatial regression model (Geographically Weighted Regression—GWR) is presented for estimating house prices according to several house characteristics
Summary
The purpose of this article is to present spatial analysis results and suggest how geovisualization can contribute to their interpretation. In order to explore the factors that contribute to property values, hedonic regression is the most common technique [1,6,15] These are regression models, in which house price is the dependent variable and a set of house characteristics the explanatory factors. Since spatial autocorrelation is detected in the error term, it is recommended that a hedonic pricing model is developed in a GIS environment, namely, a geographically weighted regression (GWR) model This method produces local regression models, allowing spatial variation of the regression coefficients [22]. In this way, it is possible to identify areas where individual regression coefficients, for example, the regression coefficient of the independent variable “size” of the house, indicate a greater or lower impact on house prices
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
