Using LiDAR-derived building meshes and post-fire assessments for 24,771 exposed structures, this paper estimates when structure-to-structure exposure drives self-sustained urban wildfire spread and identifies at-risk communities across California.

Design appraisal is represented by a novel data-driven approach that predicts prices of procedurally generated building designs, using learned estimates of financial preference of design performance.

Evaluating visual landscape quality at large-scale has remained challenging. In this paper we have introduced a spatial machine learning approach to generate an income-derived building-specific measure of Visual Capital.