Export & Visualization¶

Export to CSV for analysis, GEXF for network visualization in Gephi.

CSV Export¶

Export all paper metadata and analyses to a CSV file:

mapper.export_to_csv("corpus_export.csv")

The CSV includes:

Paper metadata (title, authors, year, journal)
AI-extracted fields (core argument, methodology, key findings)
Citation metrics (if fetched)
Concept associations

GEXF Export (Gephi)¶

Export your corpus as a network graph for visualization in Gephi:

from literature_mapper.viz import export_to_gexf

export_to_gexf(
    corpus_path=str(CORPUS_PATH),
    output_path="network_semantic.gexf",
    threshold=0.05,
    mode='semantic'
)

Export Modes¶

Mode	Description	Best For
`semantic`	Full knowledge graph	Understanding argument structure
`authors`	Co-authorship network	Research communities
`concepts`	Topic co-occurrence	Topic landscape
`river`	Concepts with temporal data	Dynamic network visualization
`similarity`	Paper similarity (Jaccard)	Finding related papers

Mode Details¶

Semantic (Default)¶

The full knowledge graph with papers, concepts, findings, methods, and limitations as nodes. Edges represent relationships like SUPPORTS, CONTRADICTS, EXTENDS.

export_to_gexf(CORPUS_PATH, "semantic.gexf", mode="semantic")

Co-Authorship¶

Network of authors weighted by shared publications:

export_to_gexf(CORPUS_PATH, "authors.gexf", mode="authors")

Use Case

Identify "invisible colleges"—clusters of researchers who frequently collaborate.

Concept Co-occurrence¶

Network of concepts that appear together in papers:

export_to_gexf(CORPUS_PATH, "concepts.gexf", mode="concepts")

River (Temporal Concepts)¶

Same as concepts, but adds a start year attribute for dynamic visualizations:

export_to_gexf(CORPUS_PATH, "river.gexf", mode="river")

Use Case

Create ThemeRiver-style visualizations in Gephi showing concept evolution.

Paper Similarity¶

Papers connected by shared concepts (Jaccard similarity):

export_to_gexf(CORPUS_PATH, "similarity.gexf", mode="similarity", threshold=0.3)

The threshold parameter controls minimum similarity for an edge (0.0–1.0).

CLI Usage¶

# Default: Semantic knowledge graph
literature-mapper viz ./my_research -o graph.gexf

# Co-authorship network
literature-mapper viz ./my_research -o authors.gexf --mode authors

# Concept co-occurrence
literature-mapper viz ./my_research -o concepts.gexf --mode concepts

# Paper similarity with threshold
literature-mapper viz ./my_research -o similar.gexf --mode similarity --threshold 0.3

Gephi Workflow¶

Open the .gexf file in Gephi
Run layout (e.g., ForceAtlas2)
Partition by node type or community
Size nodes by degree or citation count
Export as PNG, SVG, or PDF