The Run Buddy recommender leverages machine learning to provide personalized athletic clothing recommendations based on current weather conditions.
The core of Run Buddy is a machine learning model (RandomForestClassifier) that predicts appropriate running attire categories based on temperature, humidity, wind speed and other factors. The system is designed to enhance runners' experiences by suggesting suitable clothing while integrating brand partnerships.
Key Features:
- Weather-based outfit prediction using machine learning
- JSON-based data management for model, brands, user preferences, and clothing interpretations
- Flask API for seamless integration with client applications
- Continuous model improvement through user feedback
The backend is built with Python, utilizing libraries such as scikit-learn for machine learning, Flask for API development, and pandas for model serialization. The system architecture includes a modular design with separate components for model management, data collection, API handling, and testing.
This project aims to demonstrate the practical application of machine learning in everyday scenarios, providing value to athletes while exploring opportunities for brand integration and personalized recommendations.
Future plans include publishing modeling data and research to Kaggle as well as Google Colab.
Contributors are welcome to help improve the model accuracy, expand the feature set, or enhance the API functionality.
- bishop-algorithms-swift-package - Examples of commonly used algorithms and data structures in Swift Package format.
- bishop-app-runbuddy-swift - Plan your next run using Generative AI. Implemented in Swift.
Individuals are welcome to use the code with commercial and open-source projects. As a courtesy, please provide attribution to waynewbishop.com. For more information, review the complete license agreement.
Have a question? Feel free to contact me.