Detailed energy consumption data is key to data-driven urban energy modeling efforts. However, privacy considerations often prevent such data to be shared directly with researchers. Here, we present an approach based on the â€œ15/15 rule,â€ used in several states in the United States, to enable energy data to be shared while protecting the information of individual customers. We do so based on a case study typical for urban energy data, where public information is combined with privacy-sensitive data. We compare two implementations, showing that our custom algorithm achieves a 1,000 times higher computational speed at only a 10% increase in information loss compared to a previously published clustering method. Our work aims to provide a mechanism to accelerate broader energy data sharing and serve as a baseline for similar efforts in different regulatory contexts, including potential future policy frameworks based on differential privacy.
Keywords energy data, urban building energy models, privacy, data sharing, clustering, algorithms