Data-driven urban building energy modeling in Satom (CH): The energy savings potential and use of available renewable energy sources.

The thesis delves deeply into innovative methodologies aimed at enriching our comprehension of urban building energy dynamics. By merging the principles of Urban Building Energy Modeling with the strength of Machine Learning (ML) techniques, the study achieves substantial advancements in evaluating potential energy savings and harnessing renewable resources within Satom. The research journey involves developing a sturdy building model, employing Geographic Information System (GIS) software to enhance modeling precision and data aggregation. Furthermore, by incorporating state-of-the-art ML algorithms like LightGBM and Random Forest through a bottom-up strategy, the study offers accurate forecasts of energy patterns and effective renewable energy utilization. Moreover, the fusion of conventional methods like Multiple Linear Regression with ML presents an all-encompassing view of energy dynamics. Additionally, this study pioneers a forward-looking trajectory spanning three decades, meticulously assessing the energy-saving potential of buildings. This initiative intricately weaves together physical attributes, energy efficiency, and socio-economic context. By designing tailored renovation scenarios and implementing meticulous selection processes, the study identifies buildings suitable for integration into the district heating network (DHN). This iterative approach systematically optimizes the network's capacity, encapsulating a pioneering strategy that harmonizes innovation, environmental concerns, and infrastructural enhancement. In sum, the current research underscores the pivotal role of data-driven techniques in refining energy consumption and offers insights to enhance energy efficiency and nurture a greener and more sustainable urban future.