Skip to content

README.md

Latest commit

 

History

History
150 lines (109 loc) · 4.42 KB

README.md

File metadata and controls

150 lines (109 loc) · 4.42 KB

Nominative Determinism Research Platform

Quantifying the relationship between names and outcomes across multiple domains.

Overview

This platform analyzes naming patterns and their correlations with real-world phenomena:

  • Hurricane nomenclature → casualty/damage prediction (phonetic harshness, memorability)
  • Magic: The Gathering cards → competitive performance (name length, phonosemantic features)
  • Cryptocurrency markets → price trajectories (naming aesthetics, brand perception)
  • Global name diversity → economic structures (middle names, dominant names, cultural patterns)

All analyses use reproducible statistical methods: Shannon entropy, Simpson diversity, phonetic feature extraction, and sentiment scoring.

Quick Start

# Install dependencies
pip install -r requirements.txt

# Run the platform
python3 app.py

# Access at random odd port (displayed on startup)
# Default: http://localhost:<random_port>

Core Features

1. Hurricane Name Analysis

  • Data: 236 Atlantic storms (1950-2024) from NOAA HURDAT2
  • Metrics: Phonetic harshness, memorability, gender coding, sentiment
  • Finding: ROC AUC 0.916 for casualty prediction using name features alone
  • Endpoint: /hurricanes

2. MTG Card Analysis

  • Data: ~25,000 Magic cards across formats
  • Metrics: Name length, syllables, phonosemantic properties
  • Finding: Shorter, harsher names correlate with competitive play
  • Endpoint: /mtg

3. Crypto Market Intelligence

  • Data: 500+ cryptocurrencies with price history
  • Metrics: Name aesthetics, brand perception, market performance
  • API: /api/crypto/* for live analysis

4. Name Diversity Study

  • Data: 2.1M U.S. names (1880-2024) + 10 countries
  • Metrics: Shannon entropy, Gini, HHI, middle-name prevalence
  • Database: name_study.duckdb (DuckDB analytics)
  • Finding: U.S. shows exceptional diversity (HHI=20); middle names amplify 30-50%

Project Structure

├── app.py                   # Flask application (2578 lines)
├── analysis/                # Research modules
│   ├── data_acquisition.py
│   ├── processing.py
│   ├── metrics.py
│   ├── country_name_linguistics.py
│   └── america_variant_analysis.py
├── collectors/              # Data ingestion
├── analyzers/               # Feature extraction
├── data/
│   ├── raw/                 # Source data
│   └── processed/           # Cleaned datasets
├── figures/                 # Publication-ready visualizations
├── templates/               # Web UI (8 pages)
└── name_study.duckdb        # Unified analytics database

Key Endpoints

Pages

  • / — Executive dashboard
  • /analysis — Statistical findings
  • /hurricanes — Hurricane nomenclature research
  • /mtg — Magic card analysis
  • /crypto/findings — Cryptocurrency research

APIs (selected)

  • /api/hurricanes/stats — Hurricane dataset statistics
  • /api/mtg/comprehensive-report — Full MTG analysis
  • /api/crypto/advanced-stats — Crypto market metrics
  • /api/analytics/* — Predictive modeling endpoints

Data Sources

  • NOAA HURDAT2: Atlantic hurricane archive (1950-2024)
  • Scryfall API: Magic: The Gathering card database
  • CoinGecko: Cryptocurrency market data
  • U.S. SSA: Baby names dataset (1880-2024)
  • Google Books Ngrams: Historical usage patterns
  • Multiple national statistics agencies: International name data

Research Outputs

All analysis produces:

  1. Parquet datasets in data/processed/
  2. Publication figures (300 DPI PNG) in figures/
  3. DuckDB tables queryable via SQL
  4. JSON APIs for programmatic access

Citation

If using this work:

Smerconish, M. (2025). Nominative Determinism Research Platform.
Hurricane analysis: ROC AUC 0.916 for name-based casualty prediction.
Name diversity: U.S. marketplace-of-names hypothesis tested across 11 countries.

Technical Details

  • Backend: Flask (Python 3.9+)
  • Database: DuckDB (analytics), SQLite (operational)
  • Analysis: pandas, numpy, scipy, nltk
  • Visualization: matplotlib, seaborn
  • Deployment: Random odd port to avoid conflicts

Contact

Independent research by Michael Smerconish
Philadelphia, PA

License: Research and educational use
Code: Fully reproducible; see analysis/ modules