StockVision Pro

Real-time market intelligence: Automated stock screening with AI-enhanced analysis, pattern recognition, and sentiment-integrated insights for data-driven investment decisions.

Comment 1

About the Project

What Inspired Us

Our project was born from a fascination with the investment philosophies of William O'Neil and Mark Minervini. Their systematic approaches to identifying high-potential growth stocks through technical analysis inspired us to bridge traditional investment wisdom with modern AI technology.

William O'Neil's investment methodology and Mark Minervini's VCP (Volatility Contraction Pattern) recognition techniques have consistently outperformed the market, but they require extensive manual analysis and pattern recognition skills. We saw an opportunity to democratise this analysis by combining their proven strategies with AI technology.

Addressing Hong Kong Investors' Needs: Many Hong Kong investors actively trade US stocks but struggle to keep up with market movements due to their busy work schedules. Our platform provides real-time analysis and alerts, enabling Hong Kong investors to stay informed about the US markets without requiring constant monitoring. We offer a Traditional Chinese interface for easy reading while maintaining English source references for transparency and credibility. Additionally, we understand the unique challenges Hong Kong investors face, including time zone differences with US markets, language barriers with English financial news, and the need for quick decision-making during limited market hours. Our system bridges these gaps by providing AI-powered analysis in their preferred language with clear source attribution.

After-hours trading optimisation: Our AI monitoring system enables Hong Kong investors to capture US market opportunities during Hong Kong business hours, providing comprehensive analysis and insights when US markets are closed, but opportunities arise through pre-market and after-hours trading sessions.

What We Learned

We built algorithms to detect VCP patterns, Cup & Handle formations, and other technical patterns based on established trading strategies. We created a relative strength rating system that compares stock performance against the S&P 500 over multiple timeframes.

We learned to integrate Gemini 2.5 Flash with traditional technical analysis to provide additional market insights. We discovered how news sentiment and market psychology can impact stock performance beyond technical indicators, and built risk assessment features that consider volatility, liquidity, and market correlation.

How We Built It

Our system follows a modular design with four main layers: Data Layer (yfinance API, News APIs, RSS Feeds, Web Scraping), Analysis Layer (Technical Analysis, Pattern Recognition), AI Layer (Gemini 2.5 Flash, Sentiment Analysis), and Presentation Layer (Flask Web Framework, Interactive Charts, Real-time Dashboard, Responsive UI).

We built O'Neil's relative strength calculations and created algorithms to detect VCP formations, Cup & Handle breakouts, support and resistance levels, and trend continuation patterns. We combined traditional technical analysis with AI insights to build a news aggregation system that leverages multiple sources while maintaining data quality.

Our technology stack includes Python, Flask, SQLite, asyncio for the backend, Gemini 2.5 Flash and OpenRouter API for AI deployment, yfinance, pandas, numpy, and beautifulsoup4 for data processing, as well as Bootstrap 5, Plotly.js, and JavaScript for the frontend.

Challenges We Faced

Data Quality and Reliability: Integrating multiple news sources with varying data quality was a challenging task. We built data validation, link verification, and fallback mechanisms to improve data accuracy across all sources.

Performance: Processing market data while maintaining good response times required careful optimisation. We implemented caching strategies and asynchronous processing to reduce analysis time significantly.

AI Integration: Combining traditional technical analysis with AI insights while keeping it understandable was complex. We created a structured 12-point analysis framework that provides valuable and understandable AI-enhanced analysis.

Pattern Recognition: Building reliable pattern recognition algorithms required extensive testing. We refined pattern detection algorithms to improve accuracy in VCP and Cup & Handle pattern detection.

System Design: Building a system that can handle data processing efficiently requires careful architecture. We utilised a modular architecture with clear separation of concerns to create a system capable of processing over 200 stocks daily.

Key Innovations

We combined traditional technical analysis with modern AI to create a system that leverages both approaches. We made a 12-point analysis framework that provides consistent investment analysis with clear weight distributions.

We built a news aggregation system that combines multiple sources while maintaining data quality. We implemented a pattern detection that operates efficiently and accurately.

Impact and Results

Our system has successfully analysed over 200 stocks daily using technical and AI analysis. We identified growth stocks using established investment methodologies, significantly reducing the initial analysis time for stock evaluations. We made analysis accessible for individual investors.

Future Vision

We envision expanding this platform to include portfolio optimisation using AI-driven asset allocation, options analysis integration for advanced trading strategies, cryptocurrency analysis using similar pattern recognition techniques, and machine learning model training on historical market data for improved predictions.

This project represents the fusion of time-tested investment wisdom with cutting-edge AI technology, making professional-grade stock analysis accessible to everyone.

Important Disclaimer: This platform is for educational and informational purposes only. All analyses, recommendations, and insights provided are not financial advice. Past performance does not guarantee future results. Investment decisions should be based on your own research and risk tolerance. Always consult with qualified financial advisors before making investment decisions. The platform and its developers are not liable for any monetary losses incurred as a result of using this information.

Built With

Share this project:

Updates