Using Energy Metering Data to Support Census 2021: A Feasibility Study

Using Energy Metering Data to Support Census 2021: A Feasibility Study is a project funded by the Office for National Statistics’ Beyond 2011 programme.

Investigators: Dr Ben Anderson, Dr Patrick James & Prof. ‘Bakr Bahaj


The possible demise of the decennial UK Census presents social, policy and commercial researchers with both a challenge and an opportunity. The challenge is to transform ‘census-taking’ by finding robust alternative methods for creating small area socio-economic indicators over time. The opportunity is to transform the very nature of the socio-economic indicators themselves using new analytic methods applied to new geo-coded datasets and to radically accelerate the temporal cycle from decennial to annual or sub-annual production. If we are no longer to be restricted to what can be asked in a Census, what kinds of social indicators might we want or be able to produce and when?

As a first step, this small scale Office for National Statistics funded project will explore the feasibility of predicting household level characteristics from two energy monitoring data sources.

The data to be used are:

  1. An energy consumption monitoring dataset held by the University of Southampton which derives from c 300 households from two case study areas in the Solent region where the Sustainable Energy Research Group is conducting ongoing trials of a range of ‘energy efficiency’ interventions as part of the ESRC/EPSRC funded ‘The Role of Community- Based Initiatives in Energy Saving’ (Energy Communities) project. Overall energy consumption data is collected every second and can be linked to repeated six-monthly survey data on household occupancy and other variables.
  2. A similar dataset collected by the University of Loughborough and archived by the UK Data Service for future research use which links consumption at one minute intervals to a baseline household occupancy and appliance ownership survey. This dataset derives from 22 dwellings observed over two years (2008-2009) and due to it’s small size is likely to be of value only for exploratory or experimental analysis.

In both cases 30 minute summaries of the data will be used to replicate the kind of data that will initially be available from the proposed national electricity smart meter roll-out.

Posted in January, 2014