Leveraging Unsupervised Learning to Develop a Typology of Residential Water Users’ Attitudes Towards Conservation


Providing adequate water supply to the growing number of urban residents will be a challenge faced by many utility managers throughout the remainder of this century. Though traditionally water managers have looked towards supply-based solutions (e.g., expanding reservoirs), recent trends indicate a shift towards demand-side management (e.g., encouraging conservation behaviors). A major part of successfully implementing demand management strategies is understanding the community-specific attitudes and beliefs that may influence uptake of conservation behaviors. Here, we present results from a study aimed at understanding these community-specific attitudes and beliefs towards water conservation. In particular, we leverage survey data from three cities in the Southwestern United States and a state-of-the-art clustering algorithm to determine seven key archetypes of water consumers. These archetypes can be used to determine demand management strategies that might have greater (or lesser) success. This study provides transferable archetypes of consumer attitudes towards water conservation, as well as a novel interdisciplinary methodology that combines social survey data with unsupervised machine learning.

Publication Type
Journal Article
Water Resources Management