Instructions

In this article we are going to bring a complete ML pipeline that produces a trained Decision Tree model about Heart Failure Prediction, using a set of tools that allow a better management of the data flow (MLFlow, Hydra, anaconda and Wandb).

Data Informations

• Heart Failure Prediction - Kaggle;

Heart Failure Prediction: Analysis and modeling. There are some factors that affects Death Event. This dataset contains person's information like age, sex, blood pressure, smoke, diabetes, ejection fraction, creatinine phosphokinase, serum creatinine, serum sodium, time and we have to predict their DEATH EVENT.

Execution Steps

The starter kit contains all the steps we have previously completed, only slightly modified to work better together.

A few notes and instructions:

• When chaining together the steps, the output artifact of a step should be the input artifact of the next one (when applicable). Also use the artifact_type options so that the final visualization of the pipeline highlights the different steps. For example, you can use raw_data for the artifact containing the downloaded data, preprocessed_data for the artifact containing the data after the preprocessing, and so on.

• For testing, set the project_name to mlops-final-project. Once you are done developing, do a production run by changing the project_name to heart_failure_classification_prod. This way the visualization of the pipeline will not contain all your trials and errors. Remember to tag the produced model export as prod (we are going to use it in the next exercise)

• When developing, you can override the parameter main.execute_steps to only execute one or more steps of the pipeline, instead of the entire pipeline. This is useful for debugging. For example, this only executes the decision_tree step:

  mlflow run . -P hydra_options="main.execute_steps='decision_tree'"

  and this executes download and preprocess:

  mlflow run . -P hydra_options="main.execute_steps='download,preprocess'"

Used tools

• Visual Studio Code;
• Jupyter;
• Google Colab;
• Anaconda - 3;
• Wandb;
• MLflow;
• Hydra.

Authors

• Kathleen Noemi
• Pedro Henrique