BiasScope

A full-stack web application for detecting gender bias in text, based on the research of Gaucher, Friesen & Kay (2011). The tool combines rule-based word matching with Claude AI for contextual analysis, providing quantitative scores and qualitative insights.

---

Project Status

✅ Implemented

• Gender bias detection using Gaucher et al. (2011) word list
• Masculine / Feminine scoring with percentage breakdown
• Claude AI integration for contextual analysis and reformulation suggestions
• Interactive dashboard with Pie and Bar charts
• Analysis history stored in PostgreSQL
• REST API with full CRUD support
• Clean Architecture (Core / Data / Services / API)

🔄 In Progress

• URL scraping — analyze job ads directly from a link
• Text highlighting — color-coded masculine/feminine words in the input

📋 Planned

• Corpus analysis — bulk upload and trend visualization across multiple texts
• Side-by-side comparison — original vs. AI-generated neutral version
• User authentication — personal analysis history per account
• Multilingual support — Romanian and other language word lists
• Export to PDF / CSV
• Docker support for easy deployment

---

Architecture

BiasAnalyzer/
├── BiasAnalyzer.Core/        # Domain models, interfaces, enums
├── BiasAnalyzer.Data/        # EF Core DbContext, repositories, seed data
├── BiasAnalyzer.Services/    # Business logic, Gaucher matching, Claude integration
├── BiasAnalyzer.API/         # ASP.NET Core REST API, controllers
├── BiasAnalyzer.Tests/       # Unit tests
└── bias-analyzer-web/        # React 19 frontend dashboard

---

Tech Stack

Layer	Technology
Backend	ASP.NET Core (.NET 10), C#
ORM	Entity Framework Core 10
Database	PostgreSQL 14
AI	Anthropic Claude 3.5 Sonnet
Frontend	React 19, Recharts, Axios
Architecture	Clean Architecture

---

Getting Started

Prerequisites

• .NET 10 SDK
• PostgreSQL 14+
• Node.js 18+
• Anthropic API Key

Backend Setup

1. Clone the repository:

   git clone https://github.com/YOUR_USERNAME/BiasAnalyzer.git
   cd BiasAnalyzer

2. Configure your local settings in BiasAnalyzer.API/appsettings.Development.json:

   {
     "ConnectionStrings": {
       "DefaultConnection": "Host=localhost;Database=biasanalyzer;Username=YOUR_USERNAME;Password=YOUR_PASSWORD"
     },
     "Claude": {
       "ApiKey": "YOUR_ANTHROPIC_API_KEY"
     }
   }

3. Apply database migrations:

   dotnet ef database update --project BiasAnalyzer.Data --startup-project BiasAnalyzer.API

4. Run the API:

   dotnet run --project BiasAnalyzer.API

   API will be available at http://localhost:5294

Frontend Setup

cd bias-analyzer-web
npm install
npm start

Frontend will be available at http://localhost:3000

---

API Reference

Method	Endpoint	Description
POST	/api/analysis	Analyze a text for gender bias
GET	/api/analysis	Retrieve all past analyses
GET	/api/analysis/{id}	Retrieve a specific analysis

---

Research Reference

Gaucher, D., Friesen, J., & Kay, A. C. (2011). Evidence that gendered wording in job advertisements exists and sustains gender inequality. Journal of Personality and Social Psychology, 101(1), 109–128. https://doi.org/10.1037/a0022530

---

License

This project is for academic and research purposes.