Cityclic is an innovative machine learning solution developed during HackEPS for the Cityclic challenge. The system predicts user actions and identifies similar procedures to enhance user experience and operational efficiency.

Cityclic leverages a massive dataset of 5 million parameters to power its predictive capabilities. The project combines advanced machine learning algorithms with natural language processing techniques to deliver accurate predictions and useful procedure matching.

• User Action Prediction: Forecasts the next likely action a user will take based on historical patterns and contextual data
• Procedure Matching: Identifies and suggests similar procedures to streamline workflows
• Language Processing: Utilizes advanced text preprocessing and embedding techniques to enhance prediction accuracy
• Interactive Web Portal: Provides an intuitive interface for users to interact with the prediction system

• Rigorous testing of multiple algorithms to select the optimal prediction model
• Creation of text embeddings for improved natural language understanding
• Parameter optimization for handling the 5 million data points effectively

• Frontend: Angular
• Backend: Python with Flask
• Data Processing: Advanced NLP techniques for text embedding
• Machine Learning: Custom algorithms for prediction and matching

Our team followed a systematic approach:

1. Data analysis and preprocessing of the 5 million parameters
2. Algorithm selection through comparative testing
3. Model training and optimization
4. Backend development with Flask to expose ML functionality
5. Frontend creation with Angular for intuitive user interaction
6. Integration and end-to-end testing

The web portal allows users to:

• View predicted next actions based on their current context
• Discover matched procedures with similar characteristics
• Interact with recommendations in a user-friendly environment

• Real-time prediction updates
• Enhanced visualization of procedure similarities
• Mobile application development
• Integration with additional enterprise systems