This dataset has been derived by CDRC from one supplied by an undisclosed Domestic Energy Provider in the UK. The dataset contains details of smart meter measures for domestic gas and electricity consumption for the year 2015 in Great Britain. Smart meters provide continuous measures of consumption of gas and electricity and are central to better understanding the underlying behaviour of domestic energy consumption. These data are recorded automatically at half an hour aggregated intervals and can be considered large in volume since each household with dual fuel smart meters annually generates around 17,500 readings. There are thus immediate advantages of use of smart meters in research with focus upon consumer behaviour and energy policy. However, while offering greater precision in understanding these factors, smart meter data also create challenges when it comes to the generalisation of temporal profiles.

The data offer an opportunity to investigate the different patterns of domestic energy consumption at high temporal granularity (48 daily counts) for user usage averages at postcode sector level in Great Britain. It may provide greater precision in understanding the differences in temporal patterns for users that may have similar energy consumption but very different consumption habits. This dataset offers the possibility of deriving postcode sector estimates of traditional socio-economic indicators from transactional data collected by these smart meters. Linking smart meter data with other sources, like Census data or the Energy Performance Certificate data, present novel opportunities for small area population analysis and area classifications that relate energy consumption behaviour with user’s life style, at postcode sector.

Content

Data are provided in respect of:
• Averages across each postcode sector for approximately 600,000 users.
• Spatially aggregated reads at 30 minute intervals for 1.1 million smart meters.
• Type of service (gas/electricity).
• Date of the reading.

Quality, Representation and Bias

Percentage of suppressed observations (NAs): 16.47%. Usage values are supressed when the total number of meters in a postcode sector is 10 or less at any point (date) and for each energy type.

The underlying individual data that was used to created this dataset presents a very low occurrence of missing values. Missing data can only be found in some of the half hour readings for less than 1% of the users. Due to their anomalous energy readings, a small proportion of users (<0.1%) are labelled as non-residential users (this information is inferred as it is impossible to identify specific users in these data).

Out of 2.32 million smart meters installed by December 2015 in GB, the Domestic Energy Provider smart meter installations account for 26% and 21% for gas and electricity smart meters respectively. The number of smart meter users per month is not constant, with rollout continuing throughout the year. Over 700,000 were installed in Q1 2015 (~60% electricity, 40% gas) followed by roughly 100,000 in each of Q2 and Q3, and 200,000 in Q4.

The 590,000 electricity smart meters, located in 8,700 postcode sectors, account for 1,200 Gwh consumption, representing only 1.1%+ of the total domestic electricity consumption in Great Britain (2015). For gas, the 480,000 users, (8,200 postcode sectors) consumed 4,000 Gwh, accounting for 1.3%+ of the total domestic gas consumption in Great Britain (2015)

The 2015 data represent the early stages of the smart meter roll out. One of the potential sources of bias is thus identified to be associated with the fact that the first properties to receive a smart meter were those with old energy meters. Another possible bias may arise from the fact that the first households to receive an installation were those customers who were more likely to be at home during the campaign: this may skew the customer representativeness slightly towards elderly and families.

The geographical distribution of meters is slightly biased towards the North West and West Midlands regions, for both electricity and gas, where almost 30% of the smart meters are installed. In contrast, Wales and North East regions are significantly underrepresented, accounting for only 8% of the total of smart meters.