ashishpatel26/Predictive_Maintenance_using_Machine-Learning_Microsoft_Casestudy ? reverse-engineered prompt

Build me a predictive maintenance notebook that takes machine telemetry, error logs, maintenance records, failure history, and machine info, then predicts the chance a machine will fail soon and which component is likely to fail.

I want the notebook to load the csv files, clean the datetime fields, explore the data, and create useful features from the hourly sensor readings like voltage, rotation, pressure, and vibration. Then build the labels from the failure data, train a multi class classification model, and show how well it works with clear evaluation results and a few charts. Please include the full workflow in a way that is easy to follow, from data prep to model training to testing and basic visualizations.

If you need to, look up current docs online for any libraries or best practices. Keep it practical and make it run end to end in the notebook with the data in the repo.