Evaluation of green space influence on housing prices using machine learning and urban visual intelligence

Green spaces are recognised for enhancing the aesthetic value and health benefits in urban environments, which, in turn, can influence housing prices. This study evaluates the impact of visible green spaces on housing prices in Lucas County, USA, employing an innovative approach that contrasts land use data (NGVI) and street view imagery (AGVI) as quantified indicators. Leveraging a Random Forest model from 2017 to 2019, we determined the contribution of green spaces to housing prices. The Analytic Hierarchy Process (AHP) was then used to score each independent variable based on its ranking performance, thereby assessing the significance of methodological differences in environmental valuation. Our findings reveal that while AGVI typically contributes more to housing price evaluations than NGVI, the primary determinants of housing prices are still the intrinsic property characteristics and socioeconomic factors, furthermore, we observed temporal variability in the effects of visible green space on housing prices. While previous research often suggested a clear link between green space and higher property values, our result indicates this relationship may be more location-dependent. Our research highlights the importance of not overestimating the economic impact of green spaces when planning urban development. Furthermore, our research underscores the necessity of adopting a diverse methodological framework when appraising environmental attributes in housing markets, considering both objective land use data and subjective visual assessments.