Mining in weathered alkaline carbonatite complexes involves the extraction of minerals in areas where geological processes have altered the original composition of the rocks. These complexes are formed by magmatic, hydrothermal, and weathering geological events, which modify the minerals present in the rocks. This geological complexity is reflected in the processing plant, which presents different behaviours in respect to the mine region that is mined. This scenario has motivated the creation of a 3D block model with geometallurgical clusters. To achieve this, four different algorithms were tested: K-Means, Hierarchical Agglomerative Clustering, Dual Space Clustering (Dsclus), and Clustering by Autocorrelation Statistics (Acclus). The first two consider only the multivariate aspects of the data, while the latter take spatial position into account. The Dsclus was the chosen method, once it proved effective in separating zones with similar metallurgical behaviours, while respecting the spatial continuity of the established clusters as well as providing more coherent results with deposit's geology. For the spatial mapping of the defined geometallurgical domains, we propose a workflow based on a well-known geostatistical framework derived from indicator kriging. This method is herein adapted from its original formulation of mapping multiple (K) categories simultaneously to a hierarchical approach, incorporating trends during the estimations in order to well reproduce the circular aspect of the deposit. The results demonstrate the potential of the applied methodologies not only to improve the understanding of the geometallurgical characteristics of the mineral deposit but also to support mine planning and optimise production processes.