PulseScope Analytics Dashboard

Full-stack financial news sentiment analysis platform with interactive dashboard powered by ML

What is PulseScope?

PulseScope is a complete end-to-end financial news analytics platform that automatically scrapes financial news, analyzes sentiment using VADER + FinBERT, discovers topics with BERTopic, detects high-impact market events, and visualizes insights through an interactive React dashboard.

Key Features

Backend (FastAPI)

• Multi-source news scraping (CoinTelegraph, CoinDesk, CryptoNews)
• Dual sentiment analysis (VADER + FinBERT transformer models)
• Automatic topic discovery (BERTopic with UMAP + HDBSCAN)
• Event impact detection (Custom scoring algorithm)
• RESTful API (17+ endpoints with full CORS support)
• PostgreSQL database (98 articles, 25 daily metrics, 12 topics)

Frontend (React + TypeScript)

• Interactive Dashboard with 5 comprehensive pages
• Real-time sentiment trends visualization
• Topic distribution analysis with meaningful labels
• High-impact events timeline with filtering
• Article explorer with search and sentiment filters
• Responsive design with smooth animations

---

Quick Start

Prerequisites

• Python 3.13+
• PostgreSQL 16+
• Node.js 18+ and npm
• pip

Backend Setup

# 1. Clone repository
git clone https://github.com/supremkc05/PulseScope.git
cd PulseScope

# 2. Setup Python environment
python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate

# 3. Install ML dependencies
pip install -r requirements.txt

# 4. Run Jupyter notebooks (in order to generate data)
jupyter notebook
# → model/scraping_data.ipynb       (Phase 1: Scrape news)
# → model/cleaning_data.ipynb       (Phase 2: Clean data)
# → model/sentiment_analysis.ipynb  (Phase 3: Analyze sentiment)
# → model/modelling_topic.ipynb     (Phase 4: Topic modeling)

# 5. Setup PostgreSQL database
createdb pulsescope

# 6. Configure backend environment
cd backend
cp .env.example .env
# Edit .env: Add DATABASE_URL, SECRET_KEY, CORS_ORIGINS

# 7. Install backend dependencies
pip install -r requirements.txt

# 8. Migrate data to database
python migrate_data.py

# 9. Start FastAPI server
uvicorn app.main:app --reload
# API runs at http://localhost:8000
# API docs at http://localhost:8000/api/v1/docs

Frontend Setup

# 1. Navigate to frontend directory
cd ui

# 2. Install dependencies
npm install

# 3. Configure environment
cp .env.example .env
# Edit .env: Set VITE_API_BASE_URL=http://localhost:8000/api/v1

# 4. Start development server
npm run dev
# Dashboard runs at http://localhost:5173

---

Project Structure

PulseScope/
├── data/
│   ├── scrap_data/              # Raw scraped data
│   ├── topic_analyzed_data.csv  # Final analyzed data
│   └── daily_impact_metrics.csv # Daily aggregations
│
├── model/                       # Jupyter notebooks (ML pipeline)
│   ├── scraping_data.ipynb      # BeautifulSoup web scraping
│   ├── cleaning_data.ipynb      # NLTK text preprocessing
│   ├── sentiment_analysis.ipynb # VADER + FinBERT
│   └── modelling_topic.ipynb    # BERTopic clustering
│
├── backend/                     # FastAPI backend
│   ├── app/
│   │   ├── main.py              # FastAPI application
│   │   ├── api/v1/              # REST endpoints
│   │   │   ├── sentiments.py    # Sentiment endpoints
│   │   │   ├── topicss.py       # Topic analysis endpoints
│   │   │   ├── enventss.py      # High-impact events
│   │   │   └── articlee.py      # Article CRUD
│   │   ├── models/              # SQLAlchemy models
│   │   └── schemas/             # Pydantic validators
│   ├── migrate_data.py          # CSV to PostgreSQL
│   └── requirements.txt
│
├── ui/                          # React frontend
│   ├── src/
│   │   ├── pages/               # Dashboard pages
│   │   │   ├── Overview.tsx     # Main dashboard
│   │   │   ├── SentimentAnalysis.tsx
│   │   │   ├── TopicAnalysis.tsx
│   │   │   ├── HighImpactEvents.tsx
│   │   │   └── ArticleExplorer.tsx
│   │   ├── components/          # Reusable components
│   │   ├── services/            # API client (axios)
│   │   └── types/               # TypeScript interfaces
│   ├── public/
│   │   └── favicon.svg          # Custom branding
│   └── package.json
│
└── requirements.txt             # ML dependencies

---

ML Pipeline

Web Scraping → Text Cleaning → Sentiment Analysis → Topic Modeling → Event Detection
(BeautifulSoup)  (NLTK)        (VADER + FinBERT)    (BERTopic)       (Impact Score)
     ↓              ↓                  ↓                 ↓                  ↓
 Raw HTML       Clean Text       Sentiment Scores    Topic Labels     Impact Metrics
                                                                            ↓
                                                                    PostgreSQL Database
                                                                            ↓
                                                                       FastAPI REST API
                                                                            ↓
                                                                    React Dashboard

Technologies Used

Backend

Component	Technology
API Framework	FastAPI 0.109.0
Database	PostgreSQL 16 + SQLAlchemy
Caching	Redis
Sentiment	VADER + FinBERT (transformer)
Topics	BERTopic (SentenceTransformers + UMAP + HDBSCAN)
Data Processing	pandas, NumPy, scikit-learn

Frontend

Component	Technology
Framework	React 18.3.1 + TypeScript 5.6.2
Build Tool	Vite 6.0.5
HTTP Client	Axios 1.7.9
Charts	Recharts 2.15.0
Styling	Tailwind CSS 3.4.17
Icons	Lucide React 0.469.0
Animations	Framer Motion 12.0.1

---

API Endpoints

Base URL: http://localhost:8000/api/v1

Sentiment Endpoints

• GET /sentiments/summary - Overall sentiment statistics
• GET /sentiments/trends - Daily sentiment trends
• GET /sentiments/distribution - Sentiment breakdown by category

Topic Endpoints

• GET /topics/distribution - Topic distribution with article counts
• GET /topics/{topic_id} - Detailed topic information with articles

Event Endpoints

• GET /events/high-impact - High-impact days with filters
• GET /events/impact-distribution - Event impact statistics

Article Endpoints

• GET /articles/ - Paginated article list with filters
• GET /articles/by-source - Article count by news source
• GET /articles/{article_id} - Individual article details

API Documentation: Available at http://localhost:8000/api/v1/docs (Swagger UI)

---

Database Schema

Articles Table (98 records)

• article_id, title, content, source, url
• published_at, scraped_at
• vader_compound, vader_pos, vader_neg, vader_neu
• finbert_pos, finbert_neg, finbert_neu
• sentiment_label, topic_id

Daily Metrics Table (25 records)

• date, total_articles
• avg_vader, avg_finbert
• sentiment_distribution (JSON)
• impact_score

---

Configuration

Backend Environment Variables (.env)

DATABASE_URL=postgresql://username:password@localhost:5432/pulsescope
SECRET_KEY=your-secret-key
CORS_ORIGINS=http://localhost:5173,http://localhost:3000

Frontend Environment Variables (.env)

VITE_API_BASE_URL=http://localhost:8000/api/v1

Author

supremkc05

• GitHub: @supremkc05

---