Large-scale Urban Building Function Mapping by Integrating Multiple Web-based Geospatial Data

Keywords: Building function, geospatial data, urban modeling
Session Type: Virtual Paper Abstract
Day: Monday
Session Start / End Time: 2/28/2022 02:00 PM (Eastern Time (US & Canada)) - 2/28/2022 03:20 PM (Eastern Time (US & Canada))
Room: Virtual 48

Abstract

Morphological (e.g., shape, size, height) and socio-economic functional (e.g., working, living, shopping) information of buildings is highly needed for urban planning and management as well as other applications such as city-scale building energy use modeling. Due to the limited availability of socio-economic geospatial data, it is more challenging to map functional building information compared to morphological building information, especially over large areas. In this study, we proposed a new framework to map functional building information in 50 U.S. cities by integrating multiple web-based geospatial data. First, web crawler and map crawler were developed to extract point of interests (POIs), roads, and land use parcels (LUPs) from Tripadvisor.com and Google Maps, respectively. Second, an unsupervised machine learning algorithm named OneClassSVM was used to identify residential buildings based on building footprint-derived landscape features. Third, the type ratio of POIs and the area ratio of LUPs were used to identify six commercial functions (i.e., hospital, hotel, school, shop, restaurant, and office) of Microsoft building footprints. The resulting building functional maps were evaluated in Des Moines, IA, and Boston, MA with overall accuracies of 94% and 93%, and kappa coefficients of 0.62 and 0.63, respectively. With the worldwide coverage of Google Maps and Tripadvisor.com, the proposed framework is transferable to other cities over the world. The data products generated from this study are of great use for quantitative city-scale urban studies, such as building energy use modeling, at the building level.