In recent years, the growing availability of data from computational sciences has highlighted its potential to provide valuable insights and enhance predictive accuracy. In the field of aerodynamics, extensive research and optimization efforts generate significant amounts of useful data, presenting opportunities to advance engineering through data-driven and datafusion models. However, the adoption of these models remains in its early stages, with best practices still emerging. Machine learning, including techniques such as neural networks, offers powerful tools for tasks like clustering, dimensionality reduction, classification, and regression. Despite this potential, preparing and processing aerodynamic and geometric data remains challenging. These processes are often complex and tailored to specific objectives, leading to diverse interpretations and implementations of data-driven approaches. Leveraging machine learning techniques, widely used in artificial intelligence and data mining, could significantly lower computational costs for aerodynamic analysis, optimization and uncertainty quantification [1, 2]. These innovative methods pave the way for more efficient and precise solutions in aerodynamic design, although challenges in data preparation and model refinement persist. This minisymposium seeks to highlight novel strategies and recent advancements in applying machine learning and data-driven approaches to aerodynamic analysis and uncertainty quantification. It emphasizes practical challenges and explores new opportunities offered by scientific machine learning, the integration of advanced machine learning techniques with scientific computing, for the development of more efficient and effective methodologies for analysis and design.