When activities connect: Sequencing, network analysis, and energy demand modelling in the United Kingdom

Eoghan McKenna, Sarah Higginson, Tom Hargreaves, Jason Chilvers, Murray Thomson

Research output: Contribution to journalArticlepeer-review

5 Citations (Scopus)
9 Downloads (Pure)


This work applies a network analysis technique to the study of real and synthetic residential activity data commonly used in activity and energy demand research.

UK Time Use Survey activity diaries are converted into network graphs of activity sequences. Differences between weekday and weekend networks are compared using network metrics: size, density, centrality and homophily. The results show that the weekday activity sequence network is smaller, less dense, more central and has lesser homophily than the weekend network.

The technique is applied to test the validation of a model of residential active occupancy in buildings that uses a first-order Markov chain technique to generate synthetic data. The results show that the synthetic data reproduces relative differences between the network metrics for weekdays and weekends but the differences between real and synthetic data are statistically significant and greater or comparable to the differences observed between real weekday and weekend data. The first-order Markov chain technique fails to capture important characteristics of the sequence network that are present in the real data.

The analysis technique presented here can be used to improve the testing and validation of such models in future, as well the comparative analysis of sets of aggregated activity data for periods of known difference in energy demand.
Original languageEnglish
Article number101572
JournalEnergy Research & Social Science
Early online date15 May 2020
Publication statusPublished - 1 Nov 2020

Cite this