Skip to content

README.md

Latest commit

 

History

History
65 lines (50 loc) · 2.63 KB

README.md

File metadata and controls

65 lines (50 loc) · 2.63 KB

Synthetic Case Note Simulation System

This directory contains the synthetic data generation engine for creating realistic but completely fictional social services case data.

🎯 Purpose

Generate synthetic datasets that mirror real-world complexity to support:

  • Testing risk flagging algorithms
  • Validating sentiment analysis workflows
  • Training pattern detection systems
  • Benchmarking AI agent performance in sda-casenote-reader

📁 Directory Structure

input-specifications/

Expert-authored YAML configuration files

  • client-profiles.yml - Demographic patterns and risk factor combinations
  • case-complexity-levels.yml - Service intensity and documentation patterns
  • writing-style-guides.yml - Caseworker documentation styles
  • project-scenarios/ - Testing configurations for specific SDA projects

generation-engine/

R scripts for synthetic data generation

  • client-generator.R - Creates client demographic profiles with realistic risk factors
  • note-generator.R - Generates case note text with authentic writing styles
  • complexity-controller.R - Orchestrates case complexity and service patterns
  • validation-framework.R - Ensures quality and realism of generated data

output-datasets/

Generated synthetic datasets

  • Export-ready datasets formatted for sda-casenote-reader integration
  • Multiple project scenarios with different characteristics
  • Quality validation reports and metrics

testing-harness/

Quality assurance and testing framework

  • Validation scripts for ensuring realistic distributions
  • Privacy protection verification (complete fictional status)
  • Integration testing with SDA analytical pipelines

📖 Implementation Details

See implementation.md for comprehensive architecture documentation including:

  • Expert-driven specification system
  • Generation pipeline workflow
  • Quality validation framework
  • Integration with SDA workflows

🚀 Quick Start

  1. Configure specifications: Edit YAML files in input-specifications/
  2. Generate data: Run scripts in generation-engine/
  3. Validate output: Check output-datasets/ for generated files
  4. Test integration: Use testing-harness/ for quality assurance

🔧 Key Features

  • Expert-Driven: Domain experts control parameters via YAML files
  • Completely Fictional: No real client data, privacy-protected
  • Alberta-Like: Realistic demographic patterns and terminology
  • SDA-Ready: Export formats compatible with analytical workflows
  • Quality Assured: Multi-level validation for realism and consistency

For detailed implementation information, see implementation.md