Wish Wang on Community Design
Wish Wang received an undergraduate research fellowship and was mentored by Xinwei Zhuang. Read about Wish's experience below!
The rise of extreme weather events and amplified demand for energy-efficient solutions underscores the imperative for a transformative approach to community design. This research project seeks to introduce a framework that leverages graph networks to visualize, model, and optimize neighborhood-scale communities. The objective is to seek a solution that would enhance energy resilience and forge a path towards performance-driven design. Thus, designing and retrofitting the neighborhood at an urban scale that is sustainable and robust in the face of environmental adversities would be the goal.
This undergraduate research contributes to the initial phase of the project in two parts: literature review and data processing. The literature review consisted of an analysis of existing literature on graph neural network construction, as well as the evaluation, usage, and background of graph neural networks. The different types of graph neural network, as well as examples of each unique type, is included in the literature review. The review also involved applications of graph neural networks in architecture, energy, and scientific applications. The data preprocessing stage encompasses cleaning multi-source real-world data. Then, the stage is followed by the construction, evaluative testing, and visualization of a graph neural network sample using the cleaned data to ascertain feasibility.