To access smart meter data, we’ll be using the Smart Energy Research Lab (SERL) dataset that contains over two years of energy consumption and contextual data for 13,300 households in Great Britain (GB). By drawing upon Nesta’s data science capabilities, we will aim to rigorously capture the subtle variations in household energy usage.

We want to develop a better understanding of the diverse ways that households use energy. To do this, we will use smart meter 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 meter data, we can uncover 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 will be 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 will start by using smart meter data to identify features that differentiate households’ energy usage. These could include 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 will then apply data science methods to these features in order to form clusters of households that use energy in similar ways, with distinct clusters showing different energy usage patterns.

Using contextual data about the households, we will identify 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 will provide insights into potential drivers of differences in energy use.

While others have undertaken similar research, few have done so using such a large dataset of GB households. Our unique contribution will be in using the profiles we identify to develop innovative interventions that could encourage a wide range of households to use energy sustainably.