Identifying the relationship between seasonal variation in residential load and socioeconomic characteristics

Zhenyu Wang, Hao Wang

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review


Smart meter data analysis can provide insights into residential electricity consumption behaviors. Seasonal variation in consumption is not well understood but yet important to utilities for energy pricing and services. This paper aims to develop a methodology to measure seasonal variations in load patterns and identify the relationship between seasonal variation and socioeconomic factors, as socioeconomic characteristics often have great explanatory power on electricity consumption behaviors. We first model the seasonal load patterns using a two-stage K-Medoids clustering and evaluate the relative entropy of the load pattern distributions between seasons. Then we develop decision tree classifiers for each season to analyze the importance of different socioeconomic characteristics factors. Taking real-world data as a case study, we find that income level is an essential factor influencing the pattern variation across all seasons. The number of children and the elderly is also a significant factor for certain seasonal changes.
Original languageEnglish
Title of host publicationProceedings of the 2021 - ACM International Conference on Systems for Energy-Efficient Built Environments
EditorsOmprakash Gnawali, Zoltan Nagy
Place of PublicationNew York NY USA
PublisherAssociation for Computing Machinery (ACM)
Number of pages4
ISBN (Electronic)9781450391146
Publication statusPublished - 2021
EventACM Conference on Embedded Systems for Energy-Efficient Buildings 2021 - Coimbra, Portugal
Duration: 17 Nov 202118 Nov 2021
Conference number: 8th


ConferenceACM Conference on Embedded Systems for Energy-Efficient Buildings 2021
Abbreviated titleBuildSys’21
Internet address

Cite this