Citation Analysis

Analyze citation networks, compute bibliometric indicators, and identify research fronts using Semantic Scholar, OpenAlex, and CrossRef data.

When to Use

• "What's the h-index of this author?"
• "Show me the citation network for this paper"
• "Identify the most influential papers in this field"
• "Map the co-authorship network in this area"
• "What are the emerging research fronts in NLP?"
• "Analyze citation trends for CRISPR papers over time"

When NOT to Use

• Finding papers by topic (use literature-search)
• Reading or summarizing papers (use scienceclaw-summarization)
• Writing papers (use paper-writing)
• Statistical analysis unrelated to citations (use statsmodels-stats)

Bibliometric Indicators

Author-Level Metrics

import numpy as np

def h_index(citations: list[int]) -> int:
    """Compute h-index from a list of citation counts."""
    sorted_c = sorted(citations, reverse=True)
    h = 0
    for i, c in enumerate(sorted_c):
        if c >= i + 1:
            h = i + 1
        else:
            break
    return h

def g_index(citations: list[int]) -> int:
    """Compute g-index: largest g such that top g papers have >= g^2 citations."""
    sorted_c = sorted(citations, reverse=True)
    cumsum = np.cumsum(sorted_c)
    g = 0
    for i in range(len(sorted_c)):
        if cumsum[i] >= (i + 1) ** 2:
            g = i + 1
    return g

def i10_index(citations: list[int]) -> int:
    """Number of papers with 10+ citations."""
    return sum(1 for c in citations if c >= 10)

Paper-Level Metrics

• Citation count: Raw count from Semantic Scholar / OpenAlex
• Field-weighted citation impact (FWCI): Citations / expected citations in field
• Percentile rank: Position relative to same-year, same-field papers
• Citation velocity: Citations per year since publication

Journal-Level Metrics