This study employs a machine learning model to estimate the heat transfer coefficient (HTC) of a sample of dwellings in two periods, before and after retrofitting measures took place as part of the Greens Home Grant (GHG). The aim is to determine whether these measures had a statistically significant effect on reducing the HTC of the dwellings, that is, increasing their thermal efficiency. University College London (UCL) has subcontracted EDF R&D UK to undertake this study using their solution, Deconstruct+, which can estimate HTC values remotely based on a dwelling’s energy consumption and weather data.