Spatial analysis of housing prices in the Athens Region, Greece

In this paper the housing prices in the Greater Athens region are analyzed employing data for the structural and locational characteristics of dwellings. The sample is derived from the total housing supply in 2017 and includes several thousand dwellings for sale available by online real estate agencies. However only houses for which the approximate location, in terms of geographic coordinates, is available are included in the sample.At first a description of the houses is provided, in terms of their structural characteristics, such as type of the dwelling, size, floor, number of bedrooms, parking etc. In addition several characteristics relative to the location of dwellings, such as distance from the closest metro station or distance from the city center are calculated in a GIS environment.Second, the influence of several factors on housing prices, for example the provision of parking or the type of the dwelling are examined, through statistical testing.Finally, regression analysis is used in order to estimate housing prices in the Greater Athens region. The explanatory factors are structural as well as locational characteristics. Due to the spatial dependency of the residuals in Ordinary Least Squares method, a spatial regression model (Geographically Weighted Regression) is also presented, which improves the accuracy of the prediction.A general conclusion is that most locational characteristics did not contribute significantly to the explanatory power of a regression model when compared to structural characteristics. However, the importance of locational factors was more evident when some critical distances were incorporated in statistical testing