Gamified Crowdsensing for Non-Intrusive Energy Disaggregation in Office Buildings

User activity significantly contributes to energy wastage in buildings, yet traditional load analysis relies on costly, intrusive, and privacy-invasive metering infrastructure. This paper proposes gamified crowdsensing as an alternative approach to energy disaggregation, where building occupants act as “human sensors” to report energy-consumption activities. We develop a challenge-based mobile application utilizing a risk-based scoring mechanism (Impact × Likelihood) to prioritize the reporting of high-intensity loads and overcome user resistance. To ensure data integrity, user claims are cross-checked via a sparse, low-cost verification infrastructure including NFC stickers, BLE beacons, and minimal smart metering. Pilot experiments involving 80 employees across office buildings in Greece and Spain demonstrate that user claims for PC and lighting usage alone can explain 65% of the variation in aggregate device energy consumption and 62% of the variance in total building energy consumption. Our findings reveal that while human-centric models are highly effective, they are sensitive to local office culture and require site-specific calibration. This approach provides a scalable, low-intrusive, and cost-effective alter-native to traditional energy disaggregation for understanding occupant-driven loads.