PR-Holistic Traffic Prediction for Smart Cities: A Full-Cycle Approach

Project Overview

This project focuses on the operationalization of machine learning models for real-time traffic prediction using the METR-LA dataset. The goal is to experiment with models, deploy them in a scalable environment, and monitor their performance in a real-world context.

Key Phases of the Project

1. Phase 1: Model Experimentation using Kubeflow
2. Phase 2: Model Deployment using Docker and Kubernetes
3. Phase 3: Model Monitoring using Evidently
4. Phase 4: Presentation of key findings and live Q&A

Learning Outcomes

• Experiment and evaluate LSTM and another selected model on the METR-LA dataset.
• Deploy the models using Docker and Kubernetes.
• Implement real-time monitoring with an Evidently dashboard.
• Effectively communicate insights via video and live presentations.

Repository Structure

• Phase1_Model_Experimentation/: Contains the files related to model experimentation.
• Phase2_Model_Deployment/: Holds Docker and Kubernetes configuration files for model deployment.
• Phase3_Model_Monitoring/: Contains monitoring setup scripts and documentation.
• Phase4_Presentation/: Files related to the project's final presentation.
• Final_Report/: Detailed final project report summarizing the work done.
• LICENSE: The project's license information.
• .gitignore: Files and directories to be ignored by Git.
• README.md: Overall project documentation (this file).

How to Use This Repository

• Clone this repository to your local machine.
• Follow the instructions in each phase's README file for specific details on setup and execution.