A machine learning approach for mining the multidimensional impact of urban form on community scale energy consumption in cities

This paper discusses the multidimensional impact that the spatial structure of urban form has on the amount of energy consumed for community building operations. Benefiting from artificial neural networks, this study is able to factor in the many spatial dimensions of urban form and explore their combined effect on community wide net energy consumption. Nineteen indices of urban form have been measured for all zip codes in San Diego and their monthly values of energy consumption have been acquired through the county's utility company, SDG and E. Inference on the resulting predictive model has been done using Shapley values showing that the most influential indices of urban form on energy con sumption are related to the compactness, passivity, shading, and diversity of a community in the context of the case study. The results of this study contribute to the larger research of this paper on adding a spatial dimension to the existing technical discourse on improving the energy performance of community microgrids.