Bibliometrics Pipeline — Technical Documentation

Complete documentation for the EM connectomics bibliometrics analysis pipeline.

Files

1. METHODOLOGY_AND_PIPELINE.md
   • Pipeline overview and data flow
   • Metric computations (PageRank, HITS, betweenness, k-core, composite score)
   • Paper role classification (10 categories)
   • Author merging strategy (17 name merges)
   • Journal club selection philosophy
   • Visualization design rationale
   • Known limitations and biases
   • QA checklist
   • Future improvements
   • Data dictionary
2. BIBLIOGRAPHY_ANALYSIS_DOCS.md
   • Dataset overview (7,503 papers, 35,641 authors)
   • Quality assurance metrics
   • Visualization descriptions
   • Paper metrics explanation
   • Author rankings methodology
3. SEMANTIC_RELEVANCE_FLAGGING.md
   • Flags in top 100 papers potentially unrelated to EM connectomics
   • Identifies authors primarily focused on fMRI/functional connectivity
   • Recommendations for filtering/metadata
   • Core EM connectomics researchers highlighted

Quick Reference

Key Statistics:

• Papers: 7,503 (with all metrics)
• Authors: 35,641 (after 17 name merges)
• Citation edges: 85,589
• Communities: 30 (Louvain detection)
• K-core range: 0–32

Top 20 People: See METHODOLOGY_AND_PIPELINE.md Section 10 (References) Top 20 Papers: See METHODOLOGY_AND_PIPELINE.md Section 10 (References) Will Gray-Roncal: Rank #65, 28 papers (merged from 4 name variants)

Recommended Journal Club Threshold: 20 (~22 papers, balanced scope)

For implementation details, see /scripts/bibliometrics/ for all code.