REAL ESTATE PRICES – URBAN SECURITY RELATIONSHIPS: SPATIAL ANALYSES AND DEPENDENCIES

The purpose of this article is to assess the possibilities for creating an analytical database to study real estate prices. To a large extent, the article presents some of the findings of a joint project with leading Bulgarian real estate agencies. Using a suitable analytical approach and standardising the information would bring substantial benefits to the dynamic Bulgarian market. Due to the lack of tools and experience, it was necessary to select an appropriate method and to apply it to the largest possible database created with the help of information from other markets. The study focused on the impact of urban security on real estate prices. On the one hand, this is a basic determinant for customers’ choice, and on the other hand, information about security rating could be used in urban planning and management. As a result, a georeferenced dataset was created with information about the characteristics of over 191 000 properties in Denver, Colorado. The application of the selected method – the Geographically Weighted Hedonic Regression – for this dataset showed a number of issues related to hardware and software restrictions of the application, the manner of data aggregation and the presence of co-linearity between indices. The application of the Geographically Weighted Principal Analysis as a means of solving the problem of co-linearity has shown other advantages such as defining the impact of various indices in smaller urban regions. Despite using data from other markets, this research has made some important conclusions regarding the definition, collection and study of data necessary for the creation of a suitable database to analyse the Bulgarian real estate market.