Search a researcher by name or ORCID to visualize their collaboration network
#
Cluster
Name
Papers
Strength
Citations
Avg Year
Country
Statistics
No data loaded
Publications per year switch to citations →
Controls
Min. shared papers5
Max links shown1000
Year rangeall
19502025
Label scale2.0×
Work Type Filter
Byline Filter
Institution Filter
Topic Clusters
Top Journals
Top Collaborators
—
Welcome to ScholarNet
A research collaboration explorer powered by OpenAlex, the free and open academic publication database.
Search any researcher by name or ORCID to instantly map their co-authorship universe.
🔍 Search
Type a researcher name or paste an ORCID iD and hit Search. ScholarNet fetches all their publications and builds a live network of collaborators — no account needed.
🖱 Interaction
Click any node to highlight it and all its connections. Right-click a node to open the author's OpenAlex profile. Drag nodes to reposition. Scroll to zoom. Click the canvas background to deselect.
🎛 Filters
By default only the top ~100 strongest collaborations are shown to keep the graph readable — the threshold is set automatically. Adjust in the sidebar:
• Min. shared papers — lower to reveal more collaborators.
• Year range & Work type — narrow by time or publication type.
• Institution — show only co-authors from specific institutions (appears after network loads).
• Topic clusters — filter by detected research community.
• Byline filter — show only papers where the searched author's listed affiliation matches their current institution, isolating collaborations built at a specific career stage.
All filters update stats, charts, and open access breakdown together.
🗺 Views
Network — co-author graph sized by shared papers. Citations — nodes sized by co-author total citations. Funders — funding agency connections. Geo — world map of collaboration countries.
★ Co-links (enabled by default)
Reveals direct connections between co-authors themselves, not just to the searched author. Uses community detection to cluster research groups visually — turning a star graph into a research community map. Toggle with the ★ Co-links button in the top-right toolbar. Purple dashed lines = co-author links. Lower Max links shown in the sidebar for cleaner cluster separation.
📡 Data Source
All publication and author data is fetched live from OpenAlex — a free, open academic graph cataloguing 474 million scholarly works. When you search a researcher, ScholarNet calls the OpenAlex REST API to retrieve their full publication list using cursor-based pagination (200 papers per request), then enriches each co-author with citation counts, institutions, and geographic data via batched author lookups.
🕸 Network Construction
Each paper's authorship list is parsed to extract co-authors. A co-author's shared paper count is the number of papers they appear on alongside the searched researcher. Link strength is a fractional weight — each paper contributes 1 / (number of co-authors), so papers with fewer authors create stronger ties. Co-authors are ranked by shared count and the top ~100 are shown by default to keep the graph readable.
📐 Graph Layout
The network is rendered using D3.js force simulation with a custom layout pipeline. Nodes are pre-positioned into cluster sectors before the simulation begins — so research communities start physically close together — then refined by charge repulsion, link attraction, and a cluster cohesion force that continuously pulls same-topic nodes toward their group centroid. This avoids the random sprawl of a naive force layout and produces spatially meaningful clusters from the first frame.
🔬 Community Detection
With Co-links enabled, direct co-author-to-co-author edges are computed from shared papers. The Louvain algorithm — a modularity-maximising community detection method — is run on this graph to assign cluster membership, grouping researchers who frequently collaborate with each other into the same colour cluster. With Co-links off, clusters fall back to the primary research topic of shared papers as reported by OpenAlex, keeping colours stable across filter changes.
🛠 Tech Stack
Visualisation: D3.js v7 + TopoJSON. Graph layout: custom D3 force simulation with sector seeding and cohesion forces. Community detection: Louvain (implemented in JavaScript). Geographic map: D3 natural-earth projection.