Predict the remaining useful life based on telemetry

This dataset comes from a fleet of a hundred machines.

Every 24h, telemetry sends volt, pressure and vibration data. We also have the age and model of the machine, if it has been maintained in the past, or if it reported errors in the last 24h.

Based on those features, we are able to predict if the remaining useful life is below or above the 20 weeks mark. Having such a prediction is great for predictive maintenance. It allows you to more accurately budget, prevent downtimes and overall save money…