Complete the form below.

Predictive maintenance is one of the top demanded applications of predictive modelling and is seen as a life-saver in asset-heavy industries such as manufacturing and aerospace due to its potential to provide significant cost savings by reducing downtime due to mechanical problems. However, using data science for this problem is much harder than it seems. The problem becomes more challenging especially when the failures rarely happen but are very costly.

By using an example based on a synthesis of multiple real-world problems, we will share what data science techniques are best equipped for this sort of analytics. In more detail, we will look into an example to predict problems caused by component failures in order to answer the question “What is the probability that a machine will fail in the near future due to a failure of a certain component”.

In this session, you will learn:

- How you can apply data science to solve predictive maintenance problems.
- What feature engineering, modelling and evaluation techniques are commonly used for these problems.
- How you can use R and Azure Notebooks to create your predictive models.

Fidan Boylu Uz

Senior Data Scientist, Microsoft