Block-Level Building Transformation Strategies for Energy Efficiency, Thermal Comfort, and Visibility Using Bayesian Multilevel Modeling

The major objective in this research is to propose building transformation strategies for energy efficiency, thermal comfort, and visibility using a Bayesian multilevel modeling approach. To address the increasing energy demands and environmental responsibility, buildings in urban areas should be transformed to be highly energy efficient while satisfying human comfort. However, multivariate relationships between variables and performance outcomes make it difficult for researchers to discern comprehensive strategies for changing building forms. In this respect, this research explores transformation strategies that can consider multiple performance in urban blocks and multiple parameters in building forms using Bayesian multilevel additive modeling. The transformation strategies are established for Kyojima, Sumida-ward, Tokyo, Japan, by analyzing 870 existing buildings. The results enable city planners, building managers, or developers to predict urban block performance based on different scenarios of building topologies and typologies. The findings can contribute to planning an optimal urban buildings' retrofitting or redevelopment for future smart and sustainable communities.