The Impact of the Application of GIS Spatial Statistics to Hedonic Price Estimation Model for House Price Determination

Automation of the property valuation has no doubt improve the accuracy of value estimation. The capabilities of GIS technology to spatially display objects, event or phenomenon in their real time have also uplift the property market. The GIS with its statistical module have enhances the house price determination taken into account the spatial relationship between the house price and the determinant factors eliminating the problem of autocorrelation and accounting for spatial interdependency of predicting factors against the hedonic (OLS) model. The study focuses on spatial statistics utilising exploratory and geographic weighted regressions in comparison to OLS regression in housing price estimation. Data for the median house price and predicting variables for the 33 Burroughs of Greater London was accessed from the database of the UK Data Service through the webpage of the London Data Store both of the geographical boundary data and statistics dataset. The study found that autocorrelation remains a problem in OLS regression with Moran’s I value of 0.1771and P=0.0139. Exploratory regression provided three (3) variables of significant contribution to house price having highest R2 (90%) and other statistics satisfied for the model fit resulting in more accurate price estimation.