Book Image

Learning Tableau 2020 - Fourth Edition

By : Joshua N. Milligan
Book Image

Learning Tableau 2020 - Fourth Edition

By: Joshua N. Milligan

Overview of this book

Learning Tableau strengthens your command on Tableau fundamentals and builds on advanced topics. The book starts by taking you through foundational principles of Tableau. We then demonstrate various types of connections and how to work with metadata. We teach you to use a wide variety of visualizations to analyze and communicate the data, and introduce you to calculations and parameters. We then take an in-depth look at level of detail (LOD) expressions and use them to solve complex data challenges. Up next, we show table calculations, how to extend and alter default visualizations, build an interactive dashboard, and master the art of telling stories with data. This Tableau book will introduce you to visual statistical analytics capabilities, create different types of visualizations and dynamic dashboards for rich user experiences. We then move on to maps and geospatial visualization, and the new Data Model capabilities introduced in Tableau 2020.2. You will further use Tableau Prep’s ability to clean and structure data and share the stories contained in your data. By the end of this book, you will be proficient in implementing the powerful features of Tableau 2020 for decision-making.
Table of Contents (19 chapters)
9
Visual Analytics – Trends, Clustering, Distributions, and Forecasting
17
Other Books You May Enjoy
18
Index

When to use a data model, joins, or blends

In one sense, every data source you create using the latest versions of Tableau will use a data model. Even data sources using one physical table will have a corresponding object in the logical layer of a data model. But when should you relate tables using the data model, when should you join them together in the physical layer, and when should you employ blending?

Most of the time, there's no single right or wrong answer. However, here are some general guidelines to help you think through when it's appropriate to use a given approach.

In general, use a data model to relate tables:

  • When joins would make correct aggregations impossible or require complex LOD expressions to get accurate results
  • When joins would result in the duplication of data
  • When you need flexibility in showing full domains of dimensions versus only values that match across relationships
  • When you are uncertain of a data...