pdefusco / MLOps_CML_DEV_Proj

Dev Project for Data Exploration and MLFlow Experiments

CML MLOps Workshop - DEV Project

This is the DEV git repository for the CML MLOps Workshop. For a more detailed expalantion of the use case and artifacts, please visit the step by step guide.

Project Deployment

If you are following the MLOps Workshop guide, this project will be automatically deployed for you when you run script 06_mlflow_apiv2.py. As an alternative, you can just deploy this project directly in CML by manually creating a project with this git url: https://github.com/pdefusco/MLOps_CML_DEV_Proj

High Level Workflow

In this part of the workshop you will develop a Classifier to predict credit card fraud. All scripts are located in the "code" folder.

1. With a CML Session with Workbench Editor run "0_data_gen.py" to create synthetic credit card data.
2. With a CML Session with JupyterLab Editor and Spark Add-On 3.2+ run through the "1_data_exploration.ipynb" notebook to familiarize yourself with the data. Run every cell without modification.
3. Using the same CML Session, run through the "2_model_building.ipynb" notebook to build a Spark Mllib classifier and manage data with Apache Iceberg. Run every cell without modification.
4. Using the same CML Session, run through the "3_register_model.ipynb" notebook to deploy the PRD project and then the classifier in it. Run every cell without modification.
5. Navigate back to the CML Home Page. You should now see a new CML Project denominated "PRD". Open it and continue from there with the instructions provided in the readme.