Modeling transaction prices of properties based on qualitative and quantitative features

Abstract

A property is a special type of commodity, which has features which can be divided into measurable – such as area and immeasurable – such as street or dis- trict. Such a division requires dual approach to model the transaction price, de- pending on the type of accepted attributes. The existing methods of testing the real estate market are divided into qualitative and quantitative methods. The first approach is based on sociological techniques, for example the relevance of various features of the property for potential buyers, and based on experience, knowledge and intuition, which allow to express opinion on the development of market phe- nomena, if they are determined by immeasurable factors. Among the quantitative methods, dominate statistical analysis and econometric models, which depend on modeling transaction prices and real estate values [2]. The methods used so far in Poland, for the real estate market analysis, are mostly based on assigning numeri- cal values to qualitative characteristics. There is also used an approach, which sep- arates the quality from quantitative traits. After the initial analysis is done, e.g. based on a ranking, the property value is corrected by quantitative methods. The aim of this study is to extend the recently used methods on the real estate market by multi-dimensional analysis, allowing comparison of the impact of sin- gle variables, as well as the evaluation of these attributes, without assigning them arbitrary numerical values. The study was carried out by the C&RT (Classification and Regression Trees), which does not require scaling the attributes, which can describe by the quality scale. This proposal extends the existing research by taking into account the quantitative characteristics and at the same time qualitative ones (no need to assign numerical values to them.) This allows the introduction of location of the premises by the chosen street, which has been ignored so far. The research was done on the basis of about 109 housing properties located in two selected districts of the City of Krakow. Those dwellings were sold between No- vember 2008 and March 2009.