Estimating housing vacancy rates at the residential neighborhood level: The case of Zhengzhou, China

Abstract

Estimating housing vacancy rates (HVR) at the residential neighborhood or even at higher spatial levels is rarely carried out due to the challenges on availability and collection of appropriate data of high spatial resolution. In this study, we introduce a framework for estimating HVR at residential neighborhood level based on selected emerging data sources: night-time light data, very high-resolution image, Open Street Map, housing data and census data. Our developed framework consists of three steps: 1) we extract residential neighborhoods as well as detailed housing information using EO-data; 2) we spatially distribute the census population into residential neighborhoods proportional to night light emissions; 3) we estimate HVR of each residential neighborhood according to the gap between its actual population and the estimated population capacity. Based on this methodology, we find the following main results for our test case of Zhengzhou, China: 1) the average HVR is estimated at 31%; 2) with rising distance to the city center the HVR is increasing.