The UK's building stock, responsible for 13\% of national greenhouse gas emissions in 2023, plays a crucial role in meeting the country's ambitious 2030 emissions reduction target. With the UK currently off-track and the building sector's emissions reductions slowing since 2014, there is an urgent need for improved energy modelling and policy development. Addressing this challenge, we introduce a novel dataset for small-neighbourhood energy modelling in England and Wales. Covering 621k postcodes, it integrates actual energy consumption data with variables spanning building characteristics, local morphology, environment, and socio-demographics. Our analysis reveals key drivers of energy usage and challenges assumptions about data intensity in predictive modelling. This work offers a foundation for providing valuable insights for urban planning and energy policy, potentially transforming approaches to small-scale energy analysis and supporting the UK's climate goals.