Work-from-home and residential energy demand: A machine learning analysis for Swiss households

Abstract

The COVID-19 pandemic has accelerated the global adoption of hybrid working strategies, with work-from-home (WFH) becoming a prevalent practice post-pandemic (PayScale, 2024). While these shifts significantly affect energy systems, there is still limited empirical evidence on how household characteristics and WFH frequency shape residential energy demand patterns, particularly in Switzerland. Switzerland's residential sector accounted for over 30% of total energy demand (2000-2022), ranking as the second-largest consumer after transportation (ODYSSEE-MURE, 2025). The post-pandemic surge in WFH adoption since March 2020 has fundamentally altered traditional consumption patterns. However, the quantitative relationship between household characteristics and WFH-associated energy consump-tion patterns in Swiss residential buildings remains poorly characterized. This study addresses these shortcomings by developing a machine learning approach to evaluate the impact of WFH on energy demand patterns of Swiss households. The study builds on questionnaire data of 2000 households from the city of Winterthur, with the aim of addressing three key research questions: 1. What sociodemographic factors make people more likely to work from home, and how do differ-ent sociodemographic factors influence the behavior of working from home? 2. What are the energy use patterns associated with working from home? 3. How do different household characteristics with different frequencies of WFH exhibit different energy use behaviors, and what impacts do these behaviors have on energy consumption? The method of this study is based on machine learning approaches including regression and clustering algorithms to quantify heterogeneous WFH energy relationships across different household types. Prac-tically, the findings provide critical insights into how evolving WFH trends impact energy systems, sup-porting evidence-based decision-making for Switzerland's Energy Strategy 2050 through identifying tar-get groups for energy efficiency interventions and grid flexibility measures. In advancing both methodo-logical frameworks and policy insights on WFH-energy systems, this study not only bridges critical knowledge gaps but also provides actionable foundations for Switzerland while offering scalable solu-tions for future sustainable energy planning worldwide.