A Digital Platform for Smart Energy Consumption in Buildings : Enabling Energy-Aware Decisions Through a Platform for Urban Heat and Building Data

Sammanfattning

This thesis presents the development of a digital platform designed to improve residential energy management and comprehension. This initiative forms a component of the DigiCityClimate collaboration, a joint effort between KTH Royal Institute of Technology, the City of Stockholm, and ElectriCITY Innovation, aimed at leveraging data and technology for climate action. The primary challenge addressed by this research is the complexity and data limitations of existing energy management tools. Many current solutions are either overly technical or provide restricted data views, hindering users’ ability to understand their building’s energy consumption patterns and identify actionable energy-saving measures. This highlights a critical need for an integrated, user-friendly tool that combines real-time sensor data, environmental and building information, and practical energy-saving advice. To mitigate this challenge, a prototype platform was developed utilizing a technology stack comprising Node.js, Express, PostgreSQL, and Vue.js. Sensor data integration was achieved through ProptechOS, while an Azure OpenAI-powered chatbot was implemented to deliver personalized energy recommendations. The chatbot’s functionality prioritizes querying a local database of predefined questions and answers, resorting to Azure OpenAI for generative responses only when no local match is found. The resulting platform offers users a clear and accessible overview of their building’s energy usage, complemented by actionable and user-friendly suggestions. While the system’s efficiency was demonstrated as a technical proofof- concept, formal user testing was not conducted. Future development could involve expanding the platform’s capabilities to cater to diverse user groups, such as housing associations or energy professionals, through the integration of more advanced features. This project underscores the potential of combining sensor data, artificial intelligence, and intuitive design to empower residents in making informed decisions. This will then potentially lead to reduced energy consumption and broader climate change mitigation efforts.