Content
This dataset, derived by the GeoDS from a UK Domestic Energy Provider, provides smart meter readings for gas and electricity consumption in Great Britain during 2015. With readings taken at 30-minute intervals, the data offers a high level of detail about consumer energy usage, supporting research into consumer behaviour and energy policy.
The dataset includes energy consumption data from approximately 600,000 users, providing averages at the postcode sector level for gas and electricity consumption. It covers 1.1 million smart meters and aggregates the readings into 30-minute intervals. The dataset can be used to explore energy consumption patterns in great temporal detail, enabling studies into differences in consumption habits across households with similar usage.
Additionally, this dataset allows for the potential derivation of socio-economic indicators based on energy consumption, offering a novel approach to small-area population analysis when linked with other data sources such as Census or Energy Performance Certificate data.
Quality, Representation and Bias
The dataset includes some suppressed data, with 16.47% of usage values hidden due to small sample sizes (fewer than 10 meters in a postcode sector). Missing data is minimal, affecting less than 1% of users. A very small proportion of users (less than 0.1%) are labelled as non-residential due to anomalous energy usage patterns.
By the end of 2015, there were 590,000 electricity and 480,000 gas smart meters installed in Great Britain. These meters represent a small fraction of the total domestic energy consumption, 1.1% for electricity and 1.3% for gas. The geographical distribution of meters is slightly skewed, with overrepresentation in the North West and West Midlands regions, where 30% of the meters are located. In contrast, regions such as Wales and the North East are underrepresented, making up only 8% of the total installations.
Given that the data reflects the early stages of the smart meter rollout, there is a potential bias in the profile of the first adopters. These early users were more likely to be elderly individuals and families who were at home during the installation campaigns, which may skew the data towards this demographic.
Usage Considerations
This dataset provides detailed insights into energy consumption patterns but should be used with awareness of its biases, particularly in relation to geographic distribution and the early stage of the smart meter rollout. Additionally, suppressed data may affect the precision of certain analyses.
