Motivated by the study of food trade relationships within the European Union and the objective of uncovering similarities among food trade market networks, we employ a Dirichlet process mixture of centered Erdős-Rényi kernels for multiple network data, as introduced in [1]. The outcomes of our analysis are easily interpretable, and the clusters we identify exhibit distinct topological properties. Our approach can be alternatively interpreted as a strategy to address the challenge of reducing the number of layers with redundant information in multiplex network data.