Flood detection is a website used to calculate the presence or probability of flooding in a particular area.
CatboostRegressor + OPTUNA : Catboost Regressor was used because it showed high success as a machine learning model.
Docker - DockerHub : Using Docker and DockerHub, we have enabled a machine learning model to run in a virtual machine environment. This allows us to run our machine learning model in any environment without having to set up the environment each time. For this purpose, I created a Dockerfile locally and included a requirements.txt file to set up the necessary environment. Then, I pushed the Docker image I created locally to a repository on DockerHub and made my Docker image available for sharing.
Google Cloud - Kubernetes : I previously transferred the Docker image I created to a cluster I set up on Google Cloud Kubernetes Engine by cloning it from GitHub. To perform the deployment, I needed to create a workload, so I created a build.yaml file for this workload. By exposing the created workload, I obtained an endpoint, thus deploying the application.
Streamlit : I coded an interface for my application using Streamlit and created a website where I can interactively obtain results.