Predictive models on imbalanced data

This is a sample of a modeling pipeline I created in Python that predicts failures of telemetric devices. The dataset is very imbalanced, i.e. the device failure rate is 0.00085, which creates challenges in the development of the models.

The final file is called ml_models_pipeline.ipynb.