OnlyTrades

XGBoost
UI

Inspiration

The unpredictable nature of stock markets and the significant impact of timely decisions inspired us to create OnlyTrades. We wanted to leverage machine learning, specifically LSTM, to identify anomalies in stock prices that signal potential buying or selling opportunities. Our goal was to make data-driven trading accessible to everyone by combining predictive analytics with an intuitive front-end interface.

What it does

OnlyTrades analyzes historical stock market data to predict anomalies in price movements using an LSTM-based model. These anomalies serve as indicators for potential buy or sell actions. The application also provides visualizations and actionable insights, allowing users to make informed trading decisions directly from the interactive Streamlit front end.

How we built it

Backend (Model Training): Used Python with Pytorch to develop an LSTM model trained on historical stock market data. Implemented anomaly detection by identifying deviations from the predicted stock price trends.

Data Processing: Cleaned and preprocessed stock data, including normalization and feature extraction. Generated time-series inputs for the LSTM model using sliding windows.

Frontend (Streamlit): Designed an interactive dashboard to display stock trends, anomalies, and buy/sell recommendations.

Challenges we ran into

Data Challenges: Handling noisy and incomplete stock market data required extensive preprocessing and careful feature engineering. Model Performance: Fine-tuning the LSTM model to balance between overfitting and underfitting for reliable anomaly detection. Real-Time Updates: Integrating live stock data with the model and frontend posed challenges in maintaining low latency and high accuracy. Frontend Optimization: Displaying large amounts of data interactively without compromising performance on Streamlit was challenging.

Accomplishments that we're proud of

Successfully implemented an LSTM-based model capable of identifying actionable stock anomalies with high precision. Created a seamless user interface using Streamlit, making complex predictions easy to interpret for users. Integrated real-time stock data visualization and actionable insights into a single platform. Bridged the gap between cutting-edge machine learning and accessible financial tools.

What we learned

That Stock market is super chaotic.

What's next for OnlyTrades

Enhanced Model: Incorporate more advanced models like Transformers or hybrid models for improved accuracy. Additional Features: Sentiment analysis based on news and social media data. Customizable alerts for anomalies and user-defined thresholds. Portfolio Management: Add tools to simulate and manage user portfolios based on predictions. Mobile Application: Develop a mobile app to provide trading insights on the go. Deployment: Deploy on cloud platforms to handle real-time predictions and scale for multiple users.