Connected Papers

Four useful starting projects

• Enter a new field: Choose a representative survey or core paper and inspect nearby clusters. Use the map to form reading questions, not to declare that the field has been exhaustively covered.
• Expand a thesis bibliography: Start from papers already known to be relevant and check nearby work. Include a paper only after assessing its connection to the thesis argument.
• Compare research traditions: Use suitable seeds from two approaches and explore their overlap. Read the actual methods to distinguish shared citations from genuinely similar assumptions.
• Find follow-on discussion: Inspect derivative work for extensions, comparisons or reviews. Track whether a later paper supports, revises or disputes the original claim.

Platforms, inputs, outputs and setup

Connected Papers runs in a browser and uses the Semantic Scholar corpus as a discovery source. Coverage depends on the available records and relationships; a missing or incomplete record can limit what the map shows.

The output is an interactive graph with associated lists and paper details. The official methodology uses co-citation and bibliographic coupling ideas, so the network should be interpreted as a guide to related literature rather than a chart of who directly cited whom.

Data, permissions and licensing

Graph history and saved-paper collections can reveal research interests. Review account and sharing choices before using the service for a confidential project or distributing a graph link.

Discovery records do not grant permission to reproduce full papers. Follow source access conditions and cite the original research when using its findings in your own writing.