Estimation of Urban Housing Vacancy Based on Daytime Housing Exterior Images—A Case Study of Guangzhou in China

The traditional methods of estimating housing vacancies rarely use daytime exterior images to estimate vacancy rates (HVR). In view this, this study proposed the idea and method urban based on images, taking Guangzhou, China as a case study. Considering residential quarters basic evaluation unit, spatial pattern its influencing factors were studied by using average nearest neighbor analysis, kernel density estimation, autocorrelation geodetector. results show that: (1) rate can be estimated which has advantage smaller research scale, simple easy operation, short time consumption, less difficulty in data acquisition. (2) Overall, Guangzhou is low core area district, followed suburban higher outer suburb, showing increasing area–urban district–suburban–outer suburb. Additionally, it obvious agglomeration characteristics, with low–low value clustered inner circle high–high (3) (<5%) are distributed area, “dual-core” pattern, while high (>50%) suburb multi-core point both have clustering characteristics. (4) factor detector that all seven an impact rate, but degree different; distance from CBD (Central Business District) strongest influence, subway accessibility weakest influence. This provides new ideas for current vacancies, not only provide reference residents purchase houses rationally, also decision-making basis planning policy formulation megacities.