Bubble2Floor is a deep learning floor plan generation model I was assigned to implement for my Machine Learning in Design course at Carnegie Mellon. The workflow utilized computer vision and machine learning to synthesize a data set of bubble diagrams and associated floor plans. The code and algorithms for the model was provided by my instructors and we were tasked with training the model and processing some of our own diagrams to gain a better understanding of the machine learning model and navigating some of the challenges in the workflow.