As I have already stated throughout this book, the idea of plotting millions of data points will never result in a data visualization of much value. In fact, although it is perhaps technically possible to query raw big data directly, it is rarely practical.

We have discussed and (hopefully) established that effective profiling and preprocessing of raw big data is an essential (required) step in any big data visualization process.

Tableau can connect directly to local and cloud data sources, as well as import data for fast in-memory processing and visualization creation. However, just as in the previous chapter, where we explored the rich capabilities provided by the open sourced D3, the correct approach to using Tableau with big data scenarios is to preprocess the raw data source into manageable data files focused on particular objectives.

Although there are many options for data profiling, preprocessing, and manipulation, a tool that is popular with many Tableau users...