Unveiling Residential Water Consumers’ Behaviour and Profiles Through Machine Learning Techniques

The continuous development of urban areas worldwide in the near future is foreseen to boost household water demand, thus placing a challenge to the distribution and supply of drinking water. Whereas several studies demonstrated the potential of customized demand management strategies to pursue water saving attitudes in the residential sector, still their effects rely on the level of understanding we have about consumers’ typical behaviours. Retrieving information on users’ behaviors at the household level, as well as their explanatory and/or causal factors, is key to spot areas towards which water saving efforts can be prioritized. This, in turn, aids the design of personalized water demand management strategies, such as education campaigns and recommendations and, coupled with monitoring programs, allows evaluating their effects in terms of behavioral change and customers’ engagement. In this work, we contribute a data-driven approach to identify and model household water users’ consumption profiles. State-of-the-art clustering methods are coupled with machine learning techniques with the aim of extracting predominant user behaviors from a set of water consumption data collected at the household scale. This allows identifying heterogeneous groups of consumers from the studied sample, as well as characterizing them with respect to several consumption features. The approach we propose in this work is validated onto a real-world household water consumption dataset, showing its potential for understanding and modeling consumers’ profiles, as well as data mining the structure of the considered community with respect to water consumption habits, ultimately informing the bottom-up collaboration between managers and customers.