Enabling data-driven design by deriving consumer appliance use from household energy data

Achieving Net Zero requires designers to have a better understanding of the product use with studies showing user behaviour, cultural norms, seasonality and product interactions concomitantly dictate energy consumption. Data on product use can support data-driven design processes that have been shown to improve the efficiency of existing products. The paper reports a method that generates data for data-driven design processes from non-intrusive load monitoring (NILM) of household energy consumption data. The method produced appliance classification accuracies of 0.9984 while reducing sample size, sampling frequency and machine learning model complexity showing potential for it to be deployed at scale across communities.