So far in the book, we've taken a practical approach to data analysis with R. With relative ease, we've been able to answer questions about particular datasets, produce models, and export visualizations. For this reason, R is an excellent choice for rapid prototyping and analytics; it is a domain-specific language designed for statistical data analysis, and it does its job well.

In the next half of the book, we will take a look at a different approach to analytics, one that is more geared toward production environments and applications. The data science pipeline of hypothesis, acquisition, cleaning and munging, analysis, modeling, visualization, and application is not a clean and linear process by any means. Moreover, when the analysis is meant to be reproducible at scale in an automated fashion, many new considerations and requirements enter into the picture. Thus, many data applications require a broader toolkit. This toolkit should still provide rapid prototyping, be generally...