Skip to content

Temporal Analysis

Track when concepts emerged, detect trends, and identify revivals in your corpus.

Concept Timeline

See when key concepts first appeared and which papers introduced them within your corpus:

timeline = mapper.get_concept_timeline(top_n=5)

print(f"{'Concept':<25} {'First':>6} {'Peak':>6} {'Papers':>7}")
print("-" * 50)
for _, row in timeline.iterrows():
    print(f"{row['concept'][:24]:<25} {row['first_year']:>6} {row['peak_year']:>6} {row['total_papers']:>7}")

Tracing a Specific Concept

# Timeline for a specific concept
specific = mapper.get_concept_timeline("homophily")

if not specific.empty:
    for _, row in specific.iterrows():
        print(f"{row['concept']}: First appeared {row['first_year']}")
        print(f"  Peak year: {row['peak_year']}")
        print(f"  Introduced by: {row['introduced_by']}")

Trend Detection

After computing temporal statistics, identify rising and declining concepts:

# First, compute temporal stats
mapper.compute_temporal_stats()

# Then view trends
rising = mapper.get_trending_concepts(direction="rising", limit=5)
declining = mapper.get_trending_concepts(direction="declining", limit=5)
Metric Description
Trend Slope Linear regression on papers/year mentioning the concept
Positive slope Concept is "Rising" (appearing in more papers over time)
Negative slope Concept is "Declining" (appearing in fewer papers over time)

Era Detection (Revivals)

Detect concepts that disappeared for a period and then reappeared—"revivals":

eras = mapper.detect_concept_eras(gap=5)  # 5-year gap threshold

for _, row in eras.iterrows():
    print(f"{row['concept']}: {row['first_era']} → gap → {row['second_era']}")

Use Case

Era detection is useful for finding forgotten methods that were later revived, or concepts that experienced a renaissance.

Scope any search or synthesis to specific time periods:

# Only retrieve evidence from 2010-2020
recent_results = mapper.search_corpus(
    query="network centrality",
    semantic=True,
    min_year=2010,
    max_year=2020,
    limit=5
)

for r in recent_results:
    print(f"{r['year']}: {r['title']}")

CLI Usage

# Compute temporal stats (run this first!)
literature-mapper temporal ./my_research

# View rising/declining concepts
literature-mapper trends ./my_research --direction rising
literature-mapper trends ./my_research --direction declining

# Analyze a specific concept's trajectory
literature-mapper trajectory "homophily" ./my_research

# Detect concept eras (revivals after gaps)
literature-mapper eras ./my_research --gap 5