Real life Manufacturing Data is not like iris or titanic. We cannot just .fit() and .score(). We need to effectively .transform() the data. In this presentation, we will look at the complexity of manufacturing data and various ways to feature engineer manufacturing data to derive meaningful variables to solve manufacturing problems. Specifically we will look at a. Methods to construct machine learning data set for different type of Manufacturing like Lot Controlled, Serial Controlled, Batch Manufacturing, Discrete Manufacturing b. Transformation method of time series data from Machine Sensor Readings c. Various binning transformations on transaction data