Automated Grouping of Medical Codes via Multiview Banded Spectral Clustering.

•A medical concept is often described in various synonyms across different EHR systems, hindering efficient data integration and knowledge discovery.•Tremendous human efforts are needed to grouping similar medical codes into clinically relevant concepts.•This article aims to develop a scalable data-driven approach by combining multiple up-to-date data sources in an unbiased manner.•The resulting groupings not only enjoy a very high alignment with existing ontologies and hierarchies, but also yield insightful findings unique to its data-driven nature.