Tableau Best-practices: How to prepare your Tableau workbook to be efficient creating your dashboard

You wonder why professionals are so productive building their dashboards?


Efficient people in dashboarding take time to carefully prepare their data source and have a great clean workbook before starting building Data Visualizations.

Here-after, a must-do list to apply if you want to be efficient creating your dashboard:


1. Import only the data you need

When you are sourcing your workbook with data, just import the tables/sheets you need. It is counter-productive to think that the more can do the less. When you will need these additional tables, you will simply come back to the data source and import the data you need, when you need it.


2. Hide/Remove all the attributes you will not use in your dashboard

Please... How come can you be productive if your workbook contains 300 attributes from which you don't even know the meaning of 90% of them? Hide/Remove those fields!

Not only it will be easier for you to look for the meaningful attributes you need for but the performance of your dashboard will be significantly better!

Just to give a rough idea: Imagine you have 15 KPIs to analyze across 30 analytical axes? This is already a lot (even too much) no? Well, it's true. For such a case, you would need a maximum of 45 attributes in your workbook.

Just think of it and remove/hide:

  • The attributes which are not meaningful for your dashboard
  • The attributes for which you do not know how to leverage them
  • The attributes where the data quality is not good enough for analysis


3. Give a business name to all remaining attributes

In Tableau, the name of your attributes will be displayed as-is in your data visualizations. Do you really want to display technical names in your dashboard? Do you really want your users to love your dashboard when they have to remember then "VBRK.KUNNR" actually stands for "Client"? Don't you think that capitalizing all the attributes the same way would make a nicer dashboard? What about removing the "_" in the attribute names?


4. Organize your attributes in folders & hierarchies

It is always easier to find attributes is they are carefully arranged in folders/hierarchies. A folder for 'Customer' attributes, another for 'Product' attributes... Same approach for Measures, Parameters...

In addition to this, leveraging folders/hierarchies will facilitate the understanding of your data source for self-service users, the structure of the data will be easier to capture.


5. Correctly define which attribute is a Dimension, which attribute is a Measure

Even being a numeric, 'Year' is definitively a dimension. Same thing for 'Order #' or equivalent.

Take all your attributes one by one: is it an axis of analysis, or is it a metric I am happy to aggregate?


Efficient people are often lazy... 


1 minute spent preparing on the data source = 5 minutes saved creating the DataViz