New journal article - City-scale residential energy consumption prediction with a multimodal approach

Yulan Sheng, Research Fellow in Urban Heat Systems is lead author on a groundbreaking study of 143,000 homes in Sheffield that presents a fast, data-light tool to identify which properties are most in need of energy-saving upgrades. Researchers found that the size of a house and the condition of its external walls are the strongest indicators of energy consumption. Using machine learning, the tool accurately predicted energy use and highlighted older, detached homes in densely built areas as top candidates for retrofitting. The findings could help cities target climate action more efficiently and avoid costly missteps in housing upgrades.

Scientific Reports, Feb 2025

City-scale residential energy consumption prediction with a multimodal approach