Building machine learning pipelines with sklearn

The sklearn pipeline package makes it easy for us to manage multiple stages of feature engineering and modeling. Performing machine learning experiments is more just than training models. It is a combination of several factors. First, you need to cleanse and transform the data. Then, you must enrich the data with feature engineering. These common tasks can be built up into a series of steps called a pipeline. When we're trying out different variants on our experiments, we can use these pipelines to train a series of very complex steps so that they become something simple and manageable that can be reused.