Volumetric data models are used to capture, store, and analyze volumetric data in fields such as medical imaging, geosciences, and engineering. In this paper, we present a computational approach and software application that systematically links attributes to 3D geometry models in Grasshopper, an algorithm editor tightly integrated into the three-dimensional (3D) modeling CAD software Rhinoceros ® (Rhino). In our approach, we use a parametric volumetric data model to assign geometric – (curvature, inclination, and material), environmental -(illumination conditions, soil type, soil volume, soil depth), and ecological (biomass, biodiversity, species) related attributes based on spatial locations or conditions to the 3D landscape and architecture models. Parametric volumetric data models enable users to scale and adjust resolution, supporting efficient training database design for Machine Learning while optimizing memory use and preserving key information. Our results show that once functions or transformations are applied to the data, new attributes can be derived (e. g., thresholding to segment regions) and that multiple attributes can be used as training dataset for advanced machine learning models. Our machine learning model helped us to classify regions within the volumetric data, linking labels or segmenting structures based on these attributes. It also helped to extract features by identifying and linking key features, for instance the importance of soil depth on a green roof and illumination conditions on a terrace to enhance biodiversity and greenery. Additionally, the volumetric data model also integrates topological relationships through its graph structure and graph segmentation processes, facilitating advanced analysis and efficient data organization. The paper contributes to the field of computational design and computer science in engineering, an emerging field in Architecture, Engineering, and Construction (AEC). As technology advances, the use of parametric volumetric data models is likely to expand, enabling more detailed representations of 3D information.