Understanding Building Energy Efficiency with Administrative and Emerging Urban Big Data by Deep Learning in Glasgow

Abstract

With buildings consuming nearly 40% of energy in developed countries, it is important to accurately estimate and understand the building energy efficiency in a city. In this research, we propose a deep learning-based multi-source data fusion framework to estimate building energy efficiency. We consider the traditional factors associated with the building energy efficiency from the energy performance certificate for 160,000 properties (30,000 buildings) in Glasgow, UK (e.g., property structural attributes and morphological attributes), as well as the Google Street View (GSV) building façade images as a complement. We compare the performance improvements between our data-fusion framework with traditional morphological attributes and image-only models. The results show that including the building façade images from GSV, the overall model accuracy increases from 79.7% to 86.8%. A further investigation and explanation of the deep learning model are conducted to understand the relationships between building features and building energy efficiency by using Shapley Additive explanations (SHAP). Our research demonstrates the potential of using multi-source data in building energy efficiency prediction to help understand building energy efficiency at the city level to help achieve the net-zero target by 2050.

High‐resolution quantification of building stock using multi‐source remote sensing imagery and deep learning

“…In the first stage, we initialize ResNet‐18 with pre‐trained weights on ImageNet (Deng et al., 2009). This initialization strategy promotes fast convergence, improves accuracy, and reduces training time (Sun et al., 2022; Xing et al., 2020). Although the pre‐trained model can extract high‐level features from the input images, the model needs to be fine‐tuned for specific prediction tasks using the grid building‐stock dataset that we created.…”

In Search of Basement Indicators from Street View Imagery Data: An Investigation of Data Sources and Analysis Strategies

“…Due to the favourable point of view and the high availability, SVI data allow capturing many features of interest described in Section 2. SVI data have been successfully used in previous works to analyse different aspects of the built environment such as architectural style, building age, and building energy efficiency [24,25].…”

Estimating Building Energy Efficiency From Street View Imagery, Aerial Imagery, and Land Surface Temperature Data

“…Using machine learning-based methods to estimate building characteristics such as energy consumption [20,23,5,22] and efficiency [12,24], photovoltaic rooftop potential [14,13] and generation [21,18], as well as property type, age, and value [10,1] has received significant research attention. In general, these studies can be further sub-divided into top-down approaches which start with estimates for a whole city or region and disaggregate them as needed and bottom-up approaches which in turn focus on individual buildings first [4].…”

“…Apart from the studies that focus on building energy consumption, [24] presents a model which uses street view imagery and tabular data such as a building's total floor area, height, and number of open fireplaces in order to estimate a building's energy efficiency on a scale from A-G, a rating scheme introduced according to the EU's directive on the energy performance of buildings (EPBD), with "A" being the most energy efficient and "G" being the least. In a real-world case study for the city of Glasgow, more than 30,000 buildings are analyzed and the model achieves an accuracy of 86.8%.…”