Residential Property Valuation: An Application of Geographically Weighted Regression (GWR)
By Dr. Tony Hernandez, Dr. Maurice Yeates and Dr. Tony Lea
This research letter is the first in a new series of technical notes on advanced research techniques and methods. The purpose of this paper is to introduce the reader to recent developments in 'spatial' regression analysis, specifically, outlining the application of geographically weighted regression (GWR) to residential property valuation. The paper is not intended to provide a 'how-to' manual, but instead aims to highlight a range of issues that are faced when generating 'local' geographically based regression models. The research letter is divided into four parts. First, the key distinction between 'global' and 'local' statistics is outlined. Second, the mechanics of GWR are introduced. The third section presents some preliminary results from GWR analysis of property values in the Greater Toronto Area (GTA). The paper concludes with some thoughts on the benefits and potential pitfalls associated with developing 'local' statistical models.