The ability of RNA-binding proteins to form complexes with other biomolecules underpins a broad range of structural properties and functions. Understanding the subcellular distribution of RNA-binding proteins and their interacting partners in the steady state and upon perturbation can therefore shed light on these aspects. Here, we present the compartmentalized RNA-Binding Protein (or coRBP) map, an experimental resource and analytical pipeline to study subcellular RNA-binding proteins through multimodal dataset integration and machine learning. Using this approach, we generate a dataset of 1,768 known and putative RNA-binding proteins distributed in a broad panel of subcellular compartments and delineate their intermolecular and intercompartmental relationships. We also establish a hierarchy of RNA-binding protein-containing complexes at multiple scales across the cell, which suggests additional functions for multiple RNA-binding proteins. Furthermore, we investigate changes in RNA-binding protein complex composition and subcellular distribution in response to C9ORF72-associated amyotrophic lateral sclerosis/frontotemporal dementia dipeptide repeats and DNA damage stress. The coRBP map provides a resource to study the roles of RNA-binding proteins in homeostasis and disease.