The paper introduces the ade4 package, which is designed for multivariate analysis in the R statistical environment. The package implements a variety of multivariate methods, following the tradition of the French school of "Analyse des Données" and based on the duality diagram. The duality diagram is a unifying mathematical tool that encompasses standard methods like principal component analysis (PCA) and more recent methods such as RLQ analysis for linking species traits and environmental variables. The paper explains the theory of the duality diagram, discusses its implementation in ade4, and presents an example using environmental data to illustrate the philosophy of the package. The example involves analyzing a table with both quantitative and qualitative variables using Hill-Smith analysis, which is a compromise between PCA and multiple correspondence analysis (MCA). The results are visualized using a biplot, highlighting the main environmental variations and providing a synthetic typology of the sites. The paper concludes by discussing the broader applicability of the duality diagram theory and the additional methods available in ade4.The paper introduces the ade4 package, which is designed for multivariate analysis in the R statistical environment. The package implements a variety of multivariate methods, following the tradition of the French school of "Analyse des Données" and based on the duality diagram. The duality diagram is a unifying mathematical tool that encompasses standard methods like principal component analysis (PCA) and more recent methods such as RLQ analysis for linking species traits and environmental variables. The paper explains the theory of the duality diagram, discusses its implementation in ade4, and presents an example using environmental data to illustrate the philosophy of the package. The example involves analyzing a table with both quantitative and qualitative variables using Hill-Smith analysis, which is a compromise between PCA and multiple correspondence analysis (MCA). The results are visualized using a biplot, highlighting the main environmental variations and providing a synthetic typology of the sites. The paper concludes by discussing the broader applicability of the duality diagram theory and the additional methods available in ade4.