GIS-MRA Techniques in Property Valuation: A Framework for Implementation

Abstract

The awareness among valuers to improve valuation analyses has induced the use of Geographic Information System (GIS) as a spatial tool. This tool has, at least, two main attractions to property valuation. First, the capability of providing an integrated database handling system, simultaneously relating spatial and attribute data of properties. Second, the capability of spatial operation and representation of these data. These functions can be performed in conjunction with multiple regression analysis (MRA) for a number of spatial analyses to improve modelling analyses. The MRA has been used as an aspatial analysis tool for real estate problems since 1920's. Characterized by its "correlation" and "prediction" features, this technique provides statistical reasoning capabilities. However, although the statistical milieu provides models that are capable of producing relatively accurate and consistent regression estimates, the technique does not possess spatial analysis capability. This capability is felt necessary because the incidence of property values are locationally distributed. Furthermore, many complex phenomena that change spatially and temporally can only be better understood by analyzing them in a visualized manner. GIS fills in this gap. With a proper framework, both GIS and MRA can be combined as a new tool of spatial reasoning in real estate analyses, particularly valuation. The core of GIS-MRA techniques is spatial-statistical analyses of property determinants and how they affect property values. This paper presents an outline of the major aspects of GIS-MRA applications in property valuation with the objective of providing a basic theoretical framework of GIS-MRA techniques, focussing on the major steps of applications in property valuation. From the discussion, it is concluded that, if the GIS-MRA techniques are carefully employed, they will add a scientific dimension to real estate analyses.