Our biases of a good analytics stack

Business intelligence and data analytics are fields that have been around for over 60 years, so clearly there are multiple approaches to building an analytics system.

Everyone will have their own preferences and biases. We do as well. Here are the biases that have shaped how we wrote this book:

  • We prefer ELT over ETL
  • We prefer using a cloud data warehouse over an on-premise data warehouse. We also prefer MPP analytics databases over Hadoop-like systems.
  • We believe data modeling is essential in an analytics setup and should not be overlooked.
  • We think that SQL based analytics will win over non-SQL based analytics.
  • We believe that analytics workflow/operations is more important than a singular focus on visualizations.

Some of these terms might not be clear to you right now, but we will clarify what each of these terms and statements mean as we go deeper into the book.

For consistency, and for a more enjoyable reading experience, we will assume that the readers are on-board with these biases for the duration of this book.

However, throughout the book we will provide additional materials as well as our own arguments on the choices we have made. We believe that presenting our ideas this way makes it easier for you to evaluate and adapt certain aspects of our approach to your own practice later.