Understanding how households use energy is essential for devising effective strategies to reduce carbon emissions. This will become increasingly important as the uptake of low-carbon heating systems such as heat pumps puts a greater demand on the electricity grid. The granularity and coverage of smart metre data means we can build a much more detailed picture of household energy consumption than ever before.

We wanted to develop a better understanding of the diverse ways that households use energy, using smart metre data to develop a set of ‘profiles’ corresponding to different patterns of energy usage. These profiles could be used to develop personalised products, services and advice - equipping households with the knowledge and tools to make more informed energy choices.

There is currently a lack of detailed knowledge about the multiple ways in which households use energy. By leveraging smart metre data, we uncovered important information about energy consumption habits and how household appliances, physical property characteristics and demographic factors influence energy usage. This information will help us design targeted energy-saving initiatives and personalised advice to promote energy-conscious behaviours, reduce wastage and ultimately lower carbon emissions.

Our findings are published openly for wider use amongst experts in the energy sector. Energy companies and distribution network operators (DNOs) could gain insights about their customers from this work, enabling them to design new tariffs and make better predictions about future energy use. Other businesses in the energy sector could develop new services and products based on our findings, and researchers could use our profiles to ensure all types of households are represented in their research.

We started by using smart metre data to identify features that differentiate households’ energy usage. These included things like the proportion of daily energy consumption that occurs in the evening, or the difference between a household’s summer and winter usage. We then applied a data science method called clustering to form groups of households that use energy in similar ways, with distinct clusters showing different energy usage patterns.

Using contextual data about the households, we then identified what households in the same cluster have in common and how clusters differ from each other - for instance, in the numbers of people living in each household and the appliances they own. This provides insights into potential drivers of differences in energy use.

You can read the insights and applications of this work in this digital report and explore the data through this interactive data dashboard.