Fetching publications...
Powered by Semantic Scholar
How This Dashboard Makes Decisions
The explorer uses public metadata. The sections below explain how it matches profiles and groups papers; you can review and override profile selections before loading.

Identity Resolution

Why: Semantic Scholar can split the same person across several author profiles.

Rule: shared ORCID is treated as high-confidence evidence and collapsed automatically. Same-name profiles with overlapping paper titles or repeated coauthors get a middle-confidence label. Name-only matches are lower confidence and expose manual override checkboxes.

Failure mode: ORCID can be missing or attached inconsistently, so distinct ORCIDs remain separate rows.

Paper Cleaning

Profiles are merged, preprints are filtered, and papers are deduplicated by DOI first, then normalized title. This keeps profile fragmentation from double-counting output.

Themes And Communities

Louvain clustering partitions the co-authorship graph; labels are inferred from paper titles, fields, methods, and geography. One-paper clusters are hidden from top-level theme chips because they usually do not represent durable research areas.

Collaboration Score

Each shared paper contributes 1 / non-profile coauthors. A two-person paper therefore counts more than one coappearance on a giant team paper, while raw joint paper count remains visible.

Collaborator Roles

Roles summarize evidence: repeated high-score ties become core collaborators, high betweenness suggests bridges, recent sparse ties become recent collaborators, and single shared papers remain one-offs. Hover or expand a row to audit the raw evidence.

Search for any researcher to map their collaboration network

Search by name or paste an ORCID · Select multiple profiles to merge fragmented records

How it works

Hover each node for technical detail.

S2 Fetch ID ORCID Disambiguate ✗ Clean G Graph L Detect T Label ✓ Dashboard

Coverage and limitations: Results depend on what Semantic Scholar has indexed, so papers that aren't in their database won't appear here. The deduplication step catches most duplicates, but it can miss some, especially for authors with inconsistent name formatting. Articles with incomplete metadata and grey literature are excluded by design.

Year
Type
Publisher
Citations
1st Author
DOI
Themes
Methods
Authors
Abstract