Related Experiment Video
Updated: Mar 18, 2026
07:12
Façade-Level Monitoring of CO2 Variability under Urban Heat Island Conditions using Low-Cost Sensor Data Loggers
Published on: December 12, 2025
250
Deciphering exterior: building energy efficiency prediction with emerging urban big data
Maoran Sun1, Ce Hou2, Qiaosi Li3
1Sustainable Design Group, University of Cambridge, Cambridge, UK.
Npj Urban Sustainability
|March 16, 2026
Summary
This study introduces a new AI method to estimate building energy efficiency using readily available data, aiding the UK's net-zero goals. Surprisingly, deprived areas show better energy efficiency, challenging assumptions.
Area of Science:
- Environmental Science
- Computer Science
- Urban Planning
Background:
- UK households account for 25% of energy consumption and carbon emissions.
- Building sector decarbonization is crucial for sustainability.
- Traditional energy efficiency assessments are resource-intensive.
Purpose of the Study:
- To develop a novel, data-driven methodology for estimating building energy efficiency.
- To leverage artificial intelligence (AI) and diverse datasets for scalable solutions.
- To support the UK's net-zero agenda through improved building energy insights.
Main Methods:
- An end-to-end multi-channel deep learning model was designed and trained.
- Utilized high-resolution thermal infrared and optical remote sensing imagery.
- Incorporated street view images, socio-economic indicators, and building morphological data.
Main Results:
- The model achieved F1 scores of 0.64 in Glasgow and 0.69 in Edinburgh.
- Demonstrated the feasibility of estimating building energy efficiency from external data.
- Identified a surprising correlation between deprived neighborhoods and better energy efficiency.
Conclusions:
- AI and widely available data offer scalable solutions for building energy efficiency.
- The findings can significantly advance the net-zero agenda.
- Challenges traditional assumptions about energy efficiency in different socio-economic contexts.
