Connectivity is the share of entities that carry the identifier that could connect them — e.g. of the creators in a repository, how many have a name identifier (ORCID). The % is occurrence-based (identified occurrences ÷ all occurrences). The bar counts distinct entities (a person or affiliation that appears in many records is one entity): green = identified in every occurrence, yellow = identified in some (these are the quick wins — a known identifier to copy to the records that are missing it), red = never identified. Click any bar to list them, color-coded to match the bars — quick wins first (a known identifier to copy to the records that are missing it), then the ones still missing an identifier.

Likely-same people, organizations, and rights statements that appear under more than one spelling — to review and reconcile. This does not change the bars above. High confidence clusters share an identifier (the same ORCID/ROR under different names); review clusters are fuzzy string matches. “Can copy” counts occurrences that could take a known identifier already present on another spelling. Acronyms vs full names (e.g. “MIT” vs the spelled-out name) only match when they share an identifier.