← Back to Projects
Stock Sentiment Analysis & Prediction App
A full-stack web application that performs real-time sentiment analysis of stocks using Reddit, Twitter, and FinViz data. Built with FastAPI and Next.js, it scrapes recent news and social media posts, processes them using FinBERT transformer models, and predicts if a stock is likely to go UP or DOWN. Features rich data visualizations, platform selection, and adjustable analysis timeframe.
Tech Stack
Core Technologies
FastAPIFinBERTSentiment AggregationNLP
Additional Technologies
Next.jsBeautifulSoupAsyncPRAWTwikitHuggingFace TransformersPyTorchChart.jsPandasScikit-learnText CleaningToken Classification
Project Highlights
Key Features
- Multi-platform scraping support
- Transformer-based sentiment prediction
- Interactive visualizations (charts & breakdowns)
- Real-time forecasting engine
- Configurable prediction thresholds
Technical Details
- Built with FastAPI, FinBERT, Sentiment Aggregation, NLP
- 11+ supporting technologies
- Category: DATA SCIENCE