This chapter presents a comprehensive method for analyzing complex neuronal data using graph theory and UMAP (Uniform Manifold Approximation and Projection), emphasizing the identification of neuronal ensembles: groups of neurons with correlated activity seen as the key elements of neuronal circuits correlated with sensorimotor, working memory, cognitive, and behavioral functions. The process is detailed from the initial data acquisition in calcium imaging experiments through the stages of data preparation, including signal binarization and the construction of a graph-based representation of the data. The core of our analysis lies in the application of community detection algorithms, such as the Louvain algorithm, to uncover the modular structure within neuronal networks, revealing how neurons group into functionally significant ensembles. Our approach aims to improve our understanding of the structural and functional connectivity of neuronal circuits that process cognitive functions like memory and decision-making. It also allows comparing control and pathological circuits, as well as pharmacological bioassays to evaluate drugs’ effects. The methodologies employed are designed to be adaptable across various neuronal datasets, providing researchers with the tools needed to explore these techniques further. This work contributes to the broader neuroscience community by offering new insights into the organizational principles of the brain and potential strategies for addressing neurological disorders. Accompanying this chapter, a GitHub repository includes all relevant codes and examples to support reproducibility and further exploration.