The aim of the project is twofold: (1) to conduct fundamental and applied research on the fields of classification and clustering, with emphasis on the analysis of environmental data; and (2) to stimulate cooperation by bringing French and Brazilian researchers/students together to exchange ideas and experiences. As there is a complementarity (clustering, meta-heuristics, meta-learning, AutoML, complexity and instance hardness measures) between the teams involved, the research proposed will be conducted from the perspective of various joint disciplines to attain more wide-ranging results. More specifically, we will address the following problems: (1) defining frameworks able to recommend time-series forecasting algorithms more suitable to the characteristics of the data; (2) defining frameworks able to recommend better pre-trained models to new data, in a transfer-learning approach; and (3) extending a novel clustering algorithm to multi-modal data while also seeking explainable results. These tasks meet the needs of our application domain, consisting of data from environmental monitoring units in the Iguaçu River basin.