Social Smart Meter: Identifying Energy Consumption Behavior in User-Generated Content

Abstract

Having a thorough understanding of energy consumption behavior is an important element of sustainability studies. Traditional sources of information about energy consumption, such as smart meter devices and surveys, can be costly to deploy, may lack contextual information or have infrequent updates. In this paper, we examine the possibility of extracting energy consumption-related information from user-generated content. More specifically, we develop a pipeline that helps identify energy-related content in Twitter posts and classify it into four categories (dwelling, food, leisure, and mobility), according to the type of activity performed. We further demonstrate a web-based application--called Social Smart Meter--that implements the proposed pipeline and enables different stakeholders to gain an insight into daily energy consumption behavior, as well as showcase it in case studies involving several world cities.