Neighbourhood and housing price : exploration of new attributes to improve property valuation in Hong Kong

Abstract

The current practices of property valuation often encounter problems like data availability and quality, and time lags in data processing raised concerns of hindering the performance of property valuation in Hong Kong. Therefore, this research goes beyond the building and locational attributes to housing prices, and starts exploring factors in the neighbourhood. The attractiveness of the nearby environment for residential properties is greatly affected by accessibility and pleasantness from social-economic, environmental and retailing aspects. While new attributes may be discovered from the unemployment rate, greenness, crowdedness, weather patterns and retail amenities, the research gap of restaurants is filled. 2 Hypotheses are proposed based on the price ranges and types of restaurants with relevance to accessibility to evaluate their correlation with the housing prices. The feasibility of using features of the neighbourhood as attributes in property valuation is studied in this research with the use of characteristics of restaurants. ArcGIS is used to visualise geo-datasets of restaurants and residential developments, spatial distributions of these properties are observed with the help of Kernel Density Estimation (KDE) and Hot Spot Analysis. Ordinary Least Square (OLS) regression and logistic regression are also performed to further validate the observations statistically. The data of restaurant details obtained by web scraping are carefully processed and transformed into the suitable form for running Hedonic Price Models. When more data of different neighbourhood factors are timely available, bigger data can be incorporated into the comprehensive models for continuous exploration of new attributes to improve property valuation in Hong Kong.