A novel approach to deriving energy citizens profiles for effective energy Policies: Clustering and statistical analyses applied to UK data

In the transformation of the energy system of the UK, the role of its citizens is becoming increasingly important through their involvement and shaping of the future energy landscape. According to the latest UK Government report, citizens consumed almost 30 % of the total 92.3 TWh in the household sector in 2023 and produced almost 20 % of the total 384.2 million tons of carbon dioxide equivalent (MtCO₂e). Identification of energy consumption habits and CO₂ footprint patterns can help to create a more comprehensive roadmap for the UK energy transition. Previous studies have classified citizens based on limited attributes, such as energy consumption, sociodemographics, or psychological characteristics to reveal meaningful patterns. This study aims to provide a more comprehensive view by clustering UK citizens based on attributes considered in previous studies, their energy consumption, and CO₂ footprint.
To this end, a machine learning model is applied to the ECHOES (“Energy CHOices supporting the Energy Union and the SET-Plan”) data set where it estimates people's energy consumption by applying a Life Cycle Assessment method; the corresponding CO₂ footprint is then calculated using published factors from the UK government for both the housing and mobility sectors. Then, the k-means clustering method is applied to identify distinct UK citizen groups. Finally, within-cluster analysis has been carried out on the clusters to compare them based on different sociodemographic attributes.
The results reveal distinguishable clusters where almost 64 % of people are associated with low energy consumption and smaller CO₂ footprint, with the remainder represented by high energy consumption and larger CO₂ footprint. Further analysis identifies three citizen profiles, where attributes such as age, energy consumption, and CO₂ footprint in the mobility sector are key differentiators. These profiles could be used to draw comprehensive narratives for energy models and carbon-neutral scenarios. Based on this information, energy models could then be built to assess more realistic decarbonization scenarios, for end-users (including policymakers) to enable them to identify optimal options within their specific context.