Data shortage for urban energy simulations? An empirical survey on data availability and enrichment methods using machine learning

Building energy simulations at district and urban scales are vital to design and operate sustainable energy systems. In many cases, these simulations rely on enrichment methods as the required detailed data on building characteristics are often unavailable. Approaches using machine learning to address this problem have already been proposed in the literature. However, research on this topic is still at an early stage and the question of whether machine learning can offer substantial solutions has not yet been answered. The goal of this work is twofold; based on an expert survey, we identify the main challenges regarding data availability for urban energy simulations.
Furthermore, we identify possibilities of machine learning methods in the field of data enrichment and city information models to offer an initial contribution in defining further research perspectives
in this domain.