In the field of building sustainability, the challenge of effectively using building data to make green choices remains a pressing concern. It’s not just a question of collecting comprehensive data on energy consumption and other environmental indicators, but also of interpreting these data in a meaningful way. This research aims to address this challenge by introducing an innovative approach based on deep generative models. This approach uses water and electricity data as an example of a dataset we can bring together to generate data-driven visual representations. The primary aim of this study is to encourage building occupants to adopt sustainable lifestyles and environmentally-friendly choices by presenting the data in an artistic format. Secondly, it involves leveraging data inference to gather building data to determine building occupancy patterns in different sections. The first steps are to collect data on water and electricity consumption, followed by a meticulous process of data inference to build a descriptive analysis that reveals the fundamental characteristics of energy consumption. At the same time, a deep neural network is trained to infer occupancy. The sustainability classifier uses characteristics such as technology adoption and energy use to classify data into sustainability categories. Subsequently, a pre-trained extended language model associates specific keywords with each sustainability category. In the final transformation, stable diffusion is used to transform the data into visually convincing images.