LeetCode Contest Rating Predictor

Project Overview

This deep learning project offers a sophisticated solution for predicting user ratings in LeetCode coding contests. It blends a Python-based FastAPI backend with a React front-end, delivering a comprehensive and user-friendly platform. Leveraging thousands of data points, the system utilizes an LSTM neural network model to analyze and predict contest ratings, representing a significant advancement in competitive coding analytics.

Key Features

• Deep Learning Model: Uses an LSTM network to accurately predict ratings from extensive contest data.
• Extensive Data Analysis: Analyzes thousands of data points to ensure precise predictions.
• Interactive Web Interface: A React-based front-end for an engaging user experience.
• Automated Data Fetching: Utilizes GraphQL for efficient data collection from LeetCode.
• Data Preprocessing: Implements advanced techniques like MinMaxScaler for data normalization.
• Backend Server: FastAPI backend for efficient data handling and response.
• Scalable Architecture: Designed to handle large datasets and complex neural network operations.

Data Fetching and Preprocessing

• LeetCode Contest Data: Automated fetching using GraphQL queries.
• Normalization and Structuring: Utilizing MinMaxScaler for data consistency.

Machine Learning Model

• LSTM Neural Network: Optimized for time-series data analysis in competitive coding scenarios.
• Training and Evaluation: In-depth training using a vast dataset for accurate prediction capabilities.

Prediction Script

• User Input Handling: Efficiently collects user data like username, ranking, and contest details.
• Rating Prediction: Employs the trained model to forecast rating changes.

Requirements

• Python 3.x: Essential for running the backend and scripts.
• Python Libraries: As listed in requirements.txt.
• Stable Internet Connection: For data fetching and web application functionality.
• Jupyter Notebook Environment: For model training using LC_Contest_Rating_Predictor.ipynb.

File Structure

• client/: Contains the React front-end application.
• LC_Contest_Rating_Predictor.ipynb: Jupyter Notebook for LSTM model training.
• data.json, usernames.json: JSON files with processed data and user information.
• model.keras: The trained deep learning model.
• scaler.save: Serialized object for data scaling.
• main.py, check.py: FastAPI backend and utility scripts.
• requirements.txt: Dependencies for Python environment.

Usage

1. Web-Based Method:
   • Start the FastAPI backend by running python main.py.
   • Navigate to the client folder and run npm start to launch the React app.
   • Interact with the web interface for data input and receive predictions.
2. Colab Method:
   • Open LC_Contest_Rating_Predictor.ipynb in Google Colab.
   • Run the notebook cells to train the model and perform predictions.

Notes

• Ensure a stable internet connection for seamless operation.
• Accuracy and performance depend on data quality and volume.
• Refer to in-code comments for detailed guidance on each component.

Conclusion

The LeetCode Contest Rating Predictor is a cutting-edge deep learning tool designed to analyze and predict user performance in coding contests. It stands out for its ability to process large-scale data and provide accurate forecasts, making it a valuable asset for competitive coders seeking to enhance their skills and strategies.