3/30/2017
How to Build a Treemap in Tableau with Selected Levels of Detail

Earlier this month I wrote a blog post about Using Treemaps to Visualize Data where I explain how and when a treemap can be useful in data visualization. Last week Andy Kriebel posted a Workout Wednesday challenge to Highlight a Treemap, which demonstrates three different techniques to highlight a segment in a treemap. This post will demonstrate how you can change the level of detail that the treemap uses and even use different levels of detail within the nodes of the treemap.

Building a treemap

Like my friend Andy, I rarely use the Show Me in Tableau. I demonstrate it in training sessions, but I don't find it useful day-to-day. Andy used the Show Me to create his treemap for his blog post, so let's start by building one without the Show Me.

The treemap is an odd chart to create in Tableau. Why? Because it does not have a pill on the Rows or Columns. Everything is done on the Marks Card. Even so, the chart is very simple to build. For this demonstration I will be using the World Indicators data set that is part of Tableau Desktop.

Build a treemap:

   Open Tableau and select the "World Indicators" data
   Move Region to Color
   Move Country to Details
   Move Population Total to Size and change the Measure to SUM
   Add Year to Filters and select 2012 (this will filter the data to a single year)

You now have a basic treemap. You can move Country to Label if you would like to have the countries listed in the treemap and you can add other fields to the Label, for example, I added the Population Total.



That was easy! It's important to understand what is happening though. Whatever level of detail is in the marks card is the level that will split the treemap. If Region(color) and Country(Details or Label) are both on the Marks Card then the treemap will show both levels in the treemap using that hierachy. Whatever field is on top will be the primary split and so on. If Country is removed then the treemap will only show Regions without a split on Country. Now we'll add some functionality to it so we can choose various levels of detail.

Building the Level of Detail

The first step to create different levels in the treemap is to create a parameter.

Step 1: Create a new parameter

   Parameter Name: Select Region
   Data Type: String
   Allowable Values: List
   Select Add from Field and select Region
   Once the values for the Regions are filled in, add All and None to the list.
   Select OK.
   Right-click on the new parameter and select Show Parameter.

Step 2: Create Calculated Field

   Calculated Field Name: Detail Level
   Formula:
   if [Select Region] = [Region] then [Country]
   elseif [Select Region] = 'All' then [Country]
   else [Region]
   END

Step 3: Use the New field Detail Level in place of Country

This new pill will now be used in place of Country. Just drag the new field on top of the Country pill. When a Region is selected from the new parameter, the level of detail will now be selected. Let's examine what the calculation is doing.

If All is selected then it will be at the lowest level of detail - Country.
If None is selected then it will be at the highest level of detail - Region.
If any of the Regions are individually selected then it will use the Country for that selection and Region for everything else.

This parameter now allows the user to select different levels in the treemap to create different comparisons. For example, when selecting "Asia" from the list, it's easy to compare the Country of India to the Region of Africa, which is now displayed as a single Region instead of Countries. In addition, putting the Sum of the Total Population on the Label will sum the Population at the different levels being aggregated (Region and Country). Click the image below to enlarge.




This approach offers additional comparisons for the user. As I outlined in my earlier blog post about Using Treemaps to Visualize Data, when there are small numbers of categories, ex. there are only six Regions, a bar chart is almost always better. However, as research has shown, a treemap offers better accuracy for leaf to non-leaf comparisons. By adding this functionality, the user can now quickly create their own leaf to non-leaf comparison at the desired level of detail. You can download an example Tableau workbook here.

I hope you find this information helpful. If you have any questions feel free to email me at Jeff@DataPlusScience.com

Jeffrey A. Shaffer
Follow on Twitter @HighVizAbility