gca-analyzer

English | 简体中文

GCA Analyzer

Introduction

GCA Analyzer is a Python package for analyzing group communication dynamics using NLP techniques and quantitative metrics. It provides comprehensive tools for understanding participation patterns, interaction dynamics, content newness, and communication density in group communications.

Features

• Multi-language Support: Built-in support for Chinese and other languages through LLM models
• Built-in Sample Data: Includes ready-to-use sample conversations for immediate testing
• Notebook Integration: Jupyter examples for quick runs in Google Colab
• Comprehensive Metrics: Analyzes group interactions through multiple dimensions
• Automated Analysis: Finds optimal analysis windows and generates detailed statistics
• Flexible Configuration: Customizable parameters for different analysis needs
• Easy Integration: Command-line interface and Python API support

Quick Start

🚀 Latest Update: Colab Support

Experience Google Colab Online

No installation required. Click the Colab badge above to run GCA Analyzer directly in your browser and get started quickly.

Installation

# Install from PyPI
pip install gca-analyzer

# For development
git clone https://github.com/etShaw-zh/gca_analyzer.git
cd gca_analyzer
pip install -e .

Basic Usage

Option 1: Use Built-in Sample Data (Recommended for First-time Users)

Start immediately with built-in sample data:

# Use built-in sample data
python -m gca_analyzer --sample-data

# Preview the sample data first
python -m gca_analyzer --sample-data --preview

# Interactive mode with sample data (recommended)
python -m gca_analyzer --interactive

Sample Data Contents:

• 42 different engineering project conversations across multiple teams
• 2,727 authentic conversation messages
• 48 different participants from collaborative engineering sessions
• Original data filtered to include only conversations with ≥30 messages for meaningful analysis

Data Source & Citation: This sample data is adapted from the Epistemic Network Analysis (ENA) Web Tool example dataset. When using this sample data in research or publications, please cite:

• Shaffer, D. W., Collier, W., & Ruis, A. R. (2016). A tutorial on epistemic network analysis: Analyzing the structure of connections in cognitive, social, and interaction data. Journal of Learning Analytics, 3(3), 9-45.
• ENA Web Tool: https://app.epistemicnetwork.org/

Option 2: Use Your Own Data

• Prepare your communication data in CSV format with required columns:

conversation_id,person_id,time,text
1A,student1,0:08,Hello teacher!
1A,teacher,0:10,Hello everyone!

• Run analysis:

  Interactive Mode:

  python -m gca_analyzer --interactive
  # or
  python -m gca_analyzer -i
  

  Command Line Mode:

  python -m gca_analyzer --data your_data.csv
  

  Advanced Options:

  python -m gca_analyzer --data your_data.csv --output results/ --model-name your-model --console-level INFO
  

Analysis Results

The analyzer generates comprehensive statistics for GCA measures:

• Participation

  • Measures relative contribution frequency
  • Negative values indicate below-average participation
  • Positive values indicate above-average participation

• Responsivity

  • Measures how well participants respond to others
  • Higher values indicate better response behavior

• Internal Cohesion

  • Measures consistency in individual contributions
  • Higher values indicate more coherent messaging

• Social Impact

  • Measures influence on group discussion
  • Higher values indicate a stronger impact on others

• Newness

  • Measures introduction of new content
  • Higher values indicate more novel contributions

• Communication Density

  • Measures information content per message
  • Higher values indicate more information-rich messages

Results are saved as CSV files in the specified output directory.

Visualizations

The analyzer provides interactive and informative visualizations:

• Radar Plots: Compare measures across participants
• Distribution Plots: Visualize measure distributions

Results are saved as interactive HTML files in the specified output directory.

Citation

If you use GCA Analyzer in your research, please cite it as follows:

@software{xiao2025gca,
  author       = {Xiao, J.},
  title        = {etShaw-zh/gca_analyzer: GCA analyzer: A python package for group communication analysis},
  version      = {v0.4.5},
  year         = {2025},
  url          = {https://doi.org/10.5281/zenodo.15906956},
  note         = {Computer software},
}