0.2 What is a data product?

We’ll begin this book by defining the topic of this class, data products. A data product is the production output of a data analysis. For example, a data analysis might build a clever machine learning algorithm. A data product embeds that algorithm in a web site so that users can input values and get predictions. Interactive analysis web sites, graphics, apps, R packages, presentations and reports are all data products. In this book we focus only on a few of these components. Mostly for space reasons, but also because our Coursera specialization covers others (like report writing).

Before beginning this book, you should be able to use R. This language will serve as the launching point for all of our data products. Fortunately, if you don’t know R, Roger Peng has a great Coursera class and LeanPub book on the subject; take and read those first. The class runs every month and both can be obtained for free.

Why R? Well for starters, it’s what I know. But, also it’s a very prevalent data analysis language. Thus, it’s convenient to build the data product in the same language as the analysis is done in. In addition, the list of tools that one needs to learn beyond R to develop data products is massive and include: HTML, Javascript, D3, Python, Amazon Web Services, and so on. In some sense, the tools we present are best thought of as prototyping tools before building a larger production endeavor. However, for many applications, they can stand alone. Shiny, in particular, is undergoing rapid adoption, development and growth.

0.3 The goal of this book

This book (and the corresponding class) has one simple goal: get you started on making data products by introducing you to some very neat tools in R. We only scratch the surface on most of these fantastic platforms, and sadly omit some important ones. It’s best to pursue this book with a simple data project in mind. So, before begining, think of a data oriented web app that you’d like to create. Try using the tools in progress to create simplified versions of your app. Hopefully by the end you’ll have a large enough toolkit to be able to learn what you need to build your app or product.