CHANGE-SCOPE

NLP-Driven Requirement Traceability and Risk Scoring System using Text Mining

---

Team: SEDA B10 | Vishwakarma Institute of Technology, Pune
Course: Software Engineering (CS2308)
Version: 1.0.0 | May 2026

---

Overview

CHANGE-SCOPE is an intelligent, web-based requirement analysis and risk assessment platform. It applies Natural Language Processing (NLP) and Text Mining techniques to Software Requirements Specification (SRS) documents to:

• Automatically extract and structure requirement statements
• Detect dependencies and build traceability matrices
• Identify change-prone and high-risk requirements
• Perform change impact analysis
• Generate analytical reports for project planning

---

Architecture

CHANGE-SCOPE/
├── backend/                  # Python FastAPI REST API
│   ├── app/
│   │   ├── main.py           # FastAPI application entry
│   │   ├── core/             # Config, DB, Security
│   │   ├── models/           # SQLAlchemy ORM models
│   │   ├── services/         # NLP Engine, Parser, Reports
│   │   └── api/routes/       # REST endpoints
│   ├── tests/                # Pytest test suite
│   └── requirements.txt
├── frontend/                 # React + TypeScript + Vite
│   ├── src/
│   │   ├── pages/            # Dashboard, Documents, Risk, etc.
│   │   ├── components/       # Reusable UI components
│   │   ├── hooks/            # Zustand auth store
│   │   └── utils/            # Axios API client
│   └── package.json
├── docs/                     # Documentation
│   ├── architecture/         # System design docs
│   └── research/             # Research paper
└── scripts/                  # Setup and seed scripts

---

Technology Stack

Layer	Technology	Purpose
Backend Framework	FastAPI (Python)	REST API, async endpoints
NLP Processing	NLTK + scikit-learn	Tokenization, TF-IDF, cosine similarity
Database	SQLite + SQLAlchemy	Async ORM, data persistence
Authentication	JWT (python-jose) + bcrypt	RBAC security
Document Parsing	pdfplumber + python-docx	.pdf, .docx, .txt support
Report Generation	fpdf2	PDF/CSV report export
Frontend	React 18 + TypeScript + Vite	SPA dashboard
UI Styling	Tailwind CSS	Utility-first design system
State Management	Zustand	Auth + client state
Data Fetching	TanStack Query	Server state + caching
Charts	Recharts	Risk visualization

---

Quick Start

Prerequisites

• Python 3.11+
• Node.js 18+
• pip and npm

1. Clone and Setup

git clone https://github.com/your-org/CHANGE-SCOPE.git
cd CHANGE-SCOPE

2. Backend Setup

cd backend

# Create virtual environment
python -m venv venv
source venv/bin/activate          # Linux/macOS
# venv\Scripts\activate           # Windows

# Install dependencies
pip install -r requirements.txt

# Download spaCy model (optional, for enhanced NLP)
python -m spacy download en_core_web_sm

# Download NLTK data
python -c "import nltk; [nltk.download(p) for p in ['punkt','stopwords','wordnet','averaged_perceptron_tagger']]"

# Start the API server
uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

The API will be available at: http://localhost:8000
Interactive docs (Swagger UI): http://localhost:8000/docs

3. Frontend Setup

cd frontend

# Install dependencies
npm install

# Start development server
npm run dev

The frontend will be available at: http://localhost:5173

4. Create Initial Admin User

cd backend
python scripts/seed_admin.py

Default credentials:

• Email: admin@changescope.com
• Password: admin123

---

Core Workflow

1. Register / Login
        ↓
2. Upload SRS Document (.pdf / .docx / .txt)
        ↓
3. System auto-parses and extracts requirements
        ↓
4. Click "Run NLP Analysis"
        ↓
5. System computes:
   • TF-IDF vectors for each requirement
   • Cosine similarity matrix (pairwise)
   • Direct and indirect dependency detection
   • Complexity and volatility scores
   • Risk scores (weighted formula)
   • Traceability links
        ↓
6. Explore outputs:
   • Requirements view (with risk levels)
   • Traceability Matrix
   • Risk Scoring Dashboard
   • Change Impact Analysis
        ↓
7. Generate & Export Reports (PDF / CSV)

---

Risk Scoring Formula

The system computes requirement risk using a weighted composite formula:

Risk(R) = 0.40 × V(R) + 0.35 × C(R) + 0.25 × D_norm(R)

Where:

• V(R) = Volatility score — linguistic uncertainty markers, modal verbs, vague quantifiers
• C(R) = Complexity score — word count, conjunctions, conditional clauses, entity density
• D_norm(R) = Normalized dependency density = min(dep_count / 10, 1.0)

Risk levels:

• High: score ≥ 0.65
• Medium: 0.35 ≤ score < 0.65
• Low: score < 0.35

---

API Endpoints

Method	Endpoint	Description
POST	/api/v1/auth/register	Register user
POST	/api/v1/auth/login	Login (JWT)
POST	/api/v1/documents/upload	Upload SRS document
POST	/api/v1/documents/{id}/parse	Parse and extract requirements
POST	/api/v1/nlp/{id}/analyze	Run full NLP pipeline
POST	/api/v1/nlp/impact/{doc_id}/{req_id}	Change impact analysis
GET	/api/v1/requirements/	List all requirements
GET	/api/v1/traceability/matrix/{id}	Traceability matrix
GET	/api/v1/risk/summary/{id}	Risk summary
POST	/api/v1/reports/{id}/generate	Generate report
GET	/api/v1/reports/download/{id}	Download report
GET	/api/v1/admin/users	List users (Admin only)
GET	/api/v1/admin/stats	System statistics

---

User Roles (RBAC)

Role	Capabilities
BusinessAnalyst	Upload SRS, review/edit requirements, run NLP analysis
ProjectManager	View risk scores, perform impact analysis, generate reports
Developer	View traceability links and requirement dependencies
QAEngineer	View traceability matrix for test coverage verification
SystemAdministrator	Manage users, roles, backup, monitor logs

---

Running Tests

cd backend
pip install pytest pytest-asyncio pytest-cov
pytest tests/ -v --cov=app --cov-report=html

Test coverage report will be generated in htmlcov/index.html.

---

Configuration

Key settings in backend/app/core/config.py (override via .env):

SECRET_KEY=your-secret-key-here
DATABASE_URL=sqlite+aiosqlite:///./changescope.db
COSINE_SIMILARITY_THRESHOLD=0.75
WEIGHT_VOLATILITY=0.40
WEIGHT_COMPLEXITY=0.35
WEIGHT_DEPENDENCY=0.25
RISK_HIGH_THRESHOLD=0.65
RISK_MEDIUM_THRESHOLD=0.35

---

Project Team

Member	Role	Institute
Varad Khedkar	NLP & Backend Lead	VIT Pune
Gautam Pise	NLP & Integration	VIT Pune
Shivaji Rathod	Backend & Database	VIT Pune
Harshal Khade	QA & Admin Module	VIT Pune

Guide: Software Engineering Faculty, VIT Pune
Client: SYG Solutions Pvt Ltd, Kalyan West, Maharashtra

---

License

This project is developed as an academic prototype for the Software Engineering course (CS2308) at Vishwakarma Institute of Technology, Pune. All intellectual property belongs to the development team unless otherwise specified.

---

CHANGE-SCOPE v1.0.0 | May 2025 | SEDA B10 | VIT Pune