RCA/CAPA Analysis Pipeline

Automated Manufacturing Insights & Defect Reduction System

This enhanced pipeline analyzes automotive complaint data to identify failure patterns, cluster recurring defects, and generate actionable manufacturing insights to improve product quality.

Key Features

• LLM-Powered Insights: Uses Qwen/Qwen2.5-3B-Instruct for executive summaries and root cause hypothesis (requires API key).
• Pattern Clustering: Uses TF-IDF + K-Means to group similar complaints automatically.
• Visual Analytics: Generates 8 professional charts for data-driven manufacturing decisions.
• Professional Reporting: Outputs a publication-ready PDF report with embedded analytics.

Quick Start

1. Install Dependencies:

   pip install -r requirements.txt

2. Configure Environment:

   • Copy .env.example to .env
   • Add your HuggingFace API Key for AI features:

     HUGGINGFACE_API_KEY=hf_your_key_here

3. Run Analysis:

   • For a quick test with sample data:

     python analysis.py --test

   • To run with real data (configure URL in .env):

     python analysis.py

Generated Artifacts

The pipeline creates a comprehensive PDF report and a suite of visualization charts in the output/ directory:

Report	Charts
RCA_Analysis_Report.pdf	Full 7-page analysis with executive summary, CAPA recommendations, and manufacturing insights.

Visualization Gallery

Severity Analysis	Component Failures

Trend Analysis	Heatmap Analysis

Complaint Clustering	Geographic Spread

Project Structure

• analysis.py: Main pipeline script
• config.py: Centralized configuration
• llm_utils.py: AI integration module (Qwen)
• clustering.py: Machine learning clustering implementation
• visualizations.py: Chart generation engine
• sample_test_data.csv: Expanded test dataset (30 records)

License

Internal Use Only - EY Automotive Safety Analysis