- Involves ensuring all stakeholders are aware of the solution
- Release model to specific stakeholder group
- Deployment is the ultimate test in validating the model is useful
- Model results needs to be traslated so the receiving stakeholders are well trained how to use and respond
- Sufficient monitor must be deployed to ensure working feedback back loop
- The process is repeated and cyclical until we have optimal setup
- Drives model refinement and peformance tuning
- The ultimate test of model is when its put into real time use in the field
- Feedback process
- Define the review process
- Refine the model
- Redeploy the improved model
- Use clear, concise and compelling story to support the case
- Support datasets with visuals and back story behind it
- Keep balance between complexity and simlicity of the story
- Use story to convey your message, so people remember and establish connection
https://hemamurhtyportfolio.wordpress.com/2021/01/22/data-science-methodology/ https://www.kaggle.com/code/xiuwenbo666/data-science-methodology