Causal discovery is a methodological approach in data science and statistics aimed at identifying causal relationships from observational data. It leverages algorithms and statistical techniques to infer causation, rather than mere correlation, which is crucial for understanding underlying mechanisms and making predictions in complex systems.