Research / Knowledge graph

A map your bibliography can explain.

Trace a connection back to authors, titles, venues, or years—and see which references your manuscript actually cites.

A living map of references Deterministic relationships

.bib

Your bibliography Entries → relationships → context

5 entries. A library taking shape. Shared authors connect A–B and B–D. Similar titles connect A–C and B–D. Citation usage adds the manuscript’s context.

5 Bibliography entries

3 Explainable connections

2 Cited in the manuscript

04 / 04

Your citations add context.

A library becomes a map.
Shared authors draw connections.
Title overlap reveals another link.
Your citations add context.

Five illustrative BibTeX entries. Lines represent shared authors or title words; filled nodes are cited in the example manuscript. The application also supports venue and year relationships. Motion illustrates the graph; metadata similarity does not establish scientific support.

01 / Inputs

Entries become nodes.

The application parses BibTeX entries and scans citation commands in the project’s TeX files. Each graph node carries the entry’s key, metadata, cited status, and number of connections.

02 / Relationships

Every edge has a reason.

Shared author, venue, year, and title-token overlap contribute configurable weights. Title similarity uses Jaccard overlap. Minimum weight and strongest-edge limits keep the graph readable.

03 / Interpretation

Related is a starting point.

Search, filters, and citation recommendations help navigate the library. Separate citation checks flag issues such as missing keys and duplicate entries. Graph connections describe metadata relationships; they do not establish scientific support.

Inside the implementation

Based on the desktop source reviewed at revision 4059a3d . File paths are relative to the application repository.

src/latex/citationAnalysis.ts Checks citation usage and library quality
src/features/graph/knowledgeGraph.ts Builds deterministic weighted relationships