Miss Emma Shelley

Research

Currently, I am working on cleaning a dataset, ResPlan, consisting of 17,000 residential floorplans, derived from real-world real estate listings, with the aim of preparing this dataset for graph-based machine learning. Each of these floor plans will be represented using a graph, which shows how different rooms and architectural features are connected to each other. These graphs will then be used for graph machine learning, to estimate the energy performance of a building, particularly how much heating and cooling it will need, with the hope that these estimations are more accurate than traditional methods.

In the long term, this will be utilised to optimise energy performance against viral disease spread in indoor spaces.  We will extend VIRIS, a web-based app that combines viral transmission, people movement, and architectural design. This app is powered by Topologicpy, advanced spatial modelling and analysis software for architecture, engineering, and construction, developed by Dr Wassim Jabi. By modelling both energy performance and disease spread using networks, we aim to optimise these to create buildings that are both energy efficient and limit the spread of diseases.