Canonical correspondence analysis (CCA) is a multivariate method used in aquatic ecology to explore relationships between biological communities and their environment. CCA extracts environmental gradients from ecological data, which are used to describe and visualize species' habitat preferences. Unlike linear methods such as partial least squares (PLS2) or redundancy analysis, CCA is better suited for unimodal relationships between species and environmental variables. The paper discusses the assumptions of CCA and focuses on interpreting ordination diagrams. It also presents advanced applications, such as ranking environmental variables and testing effects, using macroinvertebrate data. The paper compares CCA with other methods like correspondence analysis, discriminant analysis, PLS2, and co-inertia analysis. An appendix introduces CCA-PLS, a new method combining CCA and PLS2. The paper highlights that ecological data are complex due to high species diversity and variability. Species often occur in only a subset of samples, making data resemble incidence data. Relationships between species and environmental variables are typically nonlinear and non-monotonic. CCA takes advantage of these characteristics by using weighted averaging of indicator species and regression methods. It provides a framework for estimating and testing the effects of environmental variables on communities, even when these effects are masked by other sources of variation. CCA is widely used in aquatic sciences, including limnology, marine biology, and palaeolimnology. It is commonly used to identify environmental gradients and determine important variables in community composition. In palaeolimnology, CCA is used as a preliminary analysis to assess whether variables influence present-day communities. While CCA can be used for palaeo-reconstruction, more specialized methods are available. CCA is also useful for studying seasonal and spatial variation in communities.Canonical correspondence analysis (CCA) is a multivariate method used in aquatic ecology to explore relationships between biological communities and their environment. CCA extracts environmental gradients from ecological data, which are used to describe and visualize species' habitat preferences. Unlike linear methods such as partial least squares (PLS2) or redundancy analysis, CCA is better suited for unimodal relationships between species and environmental variables. The paper discusses the assumptions of CCA and focuses on interpreting ordination diagrams. It also presents advanced applications, such as ranking environmental variables and testing effects, using macroinvertebrate data. The paper compares CCA with other methods like correspondence analysis, discriminant analysis, PLS2, and co-inertia analysis. An appendix introduces CCA-PLS, a new method combining CCA and PLS2. The paper highlights that ecological data are complex due to high species diversity and variability. Species often occur in only a subset of samples, making data resemble incidence data. Relationships between species and environmental variables are typically nonlinear and non-monotonic. CCA takes advantage of these characteristics by using weighted averaging of indicator species and regression methods. It provides a framework for estimating and testing the effects of environmental variables on communities, even when these effects are masked by other sources of variation. CCA is widely used in aquatic sciences, including limnology, marine biology, and palaeolimnology. It is commonly used to identify environmental gradients and determine important variables in community composition. In palaeolimnology, CCA is used as a preliminary analysis to assess whether variables influence present-day communities. While CCA can be used for palaeo-reconstruction, more specialized methods are available. CCA is also useful for studying seasonal and spatial variation in communities.