The first truly hands-on technical phase of the data science process is actually a combination of related tasks known as extract, transform, and load (ETL). This is where you, the data science practitioner, start to mold and shape the data so that it can be as useful as possible for the later steps in the data science process. In this lesson, you'll go through each ETL task in order.

In the previous lesson, you conducted extract, transform, and load (ETL) to ensure your data was ready for the next phase of the data science process: analysis. In some cases, an analysis of the data may be the actual final goal of the project, or it may be an important intermediary step on the road to machine learning. In either case, analyzing your data using various techniques will help you obtain useful insights into that data and what it represents. It'll also give you a better understanding of how the data needs to undergo more processing to prepare it for machine learning.

Now that your data has undergone a great deal of preparation, it's time to start looking at developing machine learning models. These models will be instrumental in achieving your business objectives because they can intelligently estimate much about the world. But before you start building these models, you need to have a firm grasp on what goes into machine learning and what it means to use machine learning to test a hypothesis.

The first type of machine learning task you'll build models for is classification. Classification has many applications across many different fields, so it's a good starting point. In this lesson, you'll train classification models, tune those models, and then evaluate them as part of a process of iterative improvement.

The next major machine learning task you'll undertake is regression. Whereas classification is about placing things in categories, regression is about estimating numbers. As with the previous lesson, in this lesson you'll train, tune, and then evaluate models that perform regression.

You've built supervised learning models using both classification and regression. But now it's time to work with unsupervised learning, where labeled data is not readily available. In this lesson, you'll implement unsupervised learning in the form of clustering models, which can group observations that share common traits. Just like before, you'll develop these models as a process of training, tuning, and evaluation.

You've put your data through the extract, transform, and load (ETL) process, conducted an analysis of the data, and developed statistical models from the data that cover the three major disciplines of machine learning: classification, regression, and clustering. But you're not done yet. Now it's time to gather your results and present them to stakeholders. After all, you undertook the data science project to achieve business goals, so you need to demonstrate that you were actually successful in doing so.

You'll wrap things up and then validate what you've learned in this course by taking an assessment.