Urbanflow: an AI-driven platform for collaborative, multidisciplinary building-level and urban-scale analysis

Traditional design and analysis of cities and buildings are fragmented across disciplines and stakeholders, such as urban planning, architecture, engineering, and quantity surveying, with teams working in silos using standalone software packages or tools. This siloed approach reduces efficiency and makes sustainable planning and design more challenging.

This paper introduces Urbanflow, a research-based, collaborative (lab spin-off) web platform that enables comprehensive analysis of urban and building science factors during early design stages. Its key modules include Energyflow, which rapidly assesses energy consumption using surrogate models trained on EnergyPlus simulations and measured energy use data, and Costflow, which estimates building costs using machine learning on real cost data. We provide validation of the surrogate models using a cluster of 38 buildings.