Due to AWS costs I've brought this infrastructure down however this project was successful and the repo can be viewed for more information.
I designed and implemented a cloud-based (AWS) workflow for generating custom weather model images from NOAA's GFS raw data using Apache Airflow to orchestrate ECS tasks. The image generation code, written in Python, is packaged in an ECR container and deployed via AWS Fargate. I leveraged Terraform to automate infrastructure provisioning, including building and pushing the container image, configuring the Fargate cluster, defining the task and service, and deploying an Airflow instance on EC2. This solution enhances automation, scalability, and operational efficiency in weather model visualization.
You can view my code repository here. You can learn more about NOAA GFS Model data here.
The images are updated four times a day.
click to enlarge - if on mobile you can flip device to see image better