Bourse Mate

💡 Inspiration

I am a high school senior, and with the increasing cost of college tuition, I am preparing to manage my expenses, including the hefty college fees. According to statistics, the U.S. student loan debt has reached a staggering $1.75 trillion (both federal and private loans) 📊. This has inspired me to find ways to make my own money on the side, whether to cover tuition or as a long-term financial strategy 💸.

🤖 What it does

Bourse Mate is a web application designed to help individuals (like me) make informed stock market predictions using advanced machine learning models 🧠📈. The platform leverages real-time stock data and AI to predict stock price movements for selected companies 🏢💹. Users can input a stock symbol, and the app will predict the next day's price, giving them a better understanding of the potential for future investments 🚀. It aims to help users navigate the financial world by making data-driven decisions 💡.

🛠️ How we built it

We built Bourse Mate using several key technologies:

1. FastAPI: For the backend, allowing for fast and efficient API calls 🚀.
2. TensorFlow: For developing and training machine learning models, specifically LSTM (Long Short-Term Memory) networks to predict stock prices 🤖.
3. Polygon API: To fetch real-time stock data from the market 📉📊.
4. HTML, CSS, and JavaScript: To create the user interface, allowing users to interact with the system through a clean and simple website 🌐.
5. Docker: For containerizing the application and ensuring it runs smoothly across different environments 🐳.

⚙️ Challenges we ran into

• Data Quality: Gathering clean, reliable, and recent stock data that would allow our machine learning model to make accurate predictions 📈.
• Model Overfitting: Fine-tuning the LSTM model to avoid overfitting and ensure generalization to new data 🤔.
• Deployment: Making sure the app works seamlessly in both local and production environments 🌍. The transition from a local Flask app to FastAPI for better scalability posed some difficulties 🛠️.
• Predictive Accuracy: Stock markets are volatile and influenced by numerous external factors, so making highly accurate predictions was a challenge 📉.

🏆 Accomplishments that we're proud of

• Successfully trained an LSTM model to predict stock prices using real-time data from the Polygon API 📊💡.
• Created a user-friendly web interface that allows users to easily access the predictions and visualize the results 🎨.
• Managed to deploy the application on cloud servers, ensuring scalability and smooth performance for multiple users ☁️.
• Integrated the model with an API, enabling users to interact with the system using simple HTTP requests 🔗.

📚 What we learned

• Machine Learning: How to work with time-series data and apply LSTM models to predict future stock prices 📈.
• API Development: Building fast and efficient APIs with FastAPI and understanding how to manage requests in real-time ⚡.
• Data Science Challenges: How difficult it can be to make accurate predictions in an inherently volatile market 🔮.
• Deployment: The complexities of deploying a full-stack application and ensuring its scalability and reliability 🌐.

🚀 What's next for Bourse Mate

• Expand Prediction Models: Incorporate more advanced models, such as reinforcement learning, to make predictions based on multiple factors, including market sentiment and external events 💡.
• Improve User Interface: Enhance the front-end with more interactive features like charts and live price tracking 📊.
• Mobile App Development: Build a mobile app version of the platform to make it more accessible to a broader audience 📱.
• Investment Portfolio Integration: Add a feature to allow users to track their portfolios and receive personalized investment advice based on their current holdings 💼.

Built With

• api
• css
• fast
• flask
• google-collab
• html
• javascript
• python
• tensor-flow