Wildfire’s destruction of homes is an increasingly serious global problem. Characterizing home hardening and defensible space at the individual structure level may reduce loss through enriched understanding of structure susceptibility. However, improved data and methods are required to accurately characterize parcel-level features at scale. Here, we present a remote sensing and statistical learning approach to representing single-structure susceptibility in Ouray County, Colorado. Using LiDAR and property records, we characterize fuel attributes in the built environment. This method uses open-source data and may be used by communities to develop policies and mitigation programs.