Was that intended? Common Tableau drag n’ drop “accidents”

I recently delivered a 3-day customer on-site Tableau desktop training course for Standard Chartered Bank in Singapore and one very fair point raised was the ‘precision’ required when dragging and dropping Dimensions and Measures into the view.

I’m sure most of you have noticed that depending on how you drag n’ drop, extremely different results are produced. A reoccurring phrase from all courses I’ve taught is “I did that by accident….”

As such, I thought I’d quickly demystify a few common drag n’ drop ‘accidents’.

 

1 – Combined Axis Charts

One of the exercises in the manual asks students to create a combined axis chart (that is putting a second measure along the same axis as the first measure, automatically triggering ‘Measure Names’ and ‘Measure Values’). In a nutshell, MN/MV serve as containers for more than one measure and re-arranging MN gives you that added flexibility in how you choose to slice n’ dice a view.

 

 

What should be done is this – trigger the double green ruler when you bring in a second measure.

RIGHT:

 

 

 

WRONG:

What often happens is….

a. People trigger the single ruler…..

 

 

 

…and instead produce a dual axis chart instead of the intended combined axis chart.

 

 

b. or create a colour gradient instead when the black box is triggered

 

 

To get a separate colour, one for Profit and one for Sales, you need to drop Measure Names to colour, either by bringing MN from the Data pane OR by dragging the auto-generated MN from Rows/Columns (holding down the CTRL key).

Holding down CTRL ensures that you’re duplicating an instance of Measure Names instead of removing it from Rows/Columns, i.e. removing the ‘partition’.

Thus the lack of ‘CTRL’ when choosing to do it from Rows/Columns…….

 

 

…..produces this result:

 

 

See what just happened? You’ve now removed the Column partition and forced the measures to be ‘stacked’ on top of one another.

 

2 –  Dual Axis Charts

This time you do need to trigger the single green ruler to produce a dual axis chart.

RIGHT:

 

 

WRONG:

 

 

One measure on each axis produces a scatter plot, which is fine, but not what was originally intended:

 

 

3 – Switching between Discrete / Continuous dates

Working with dates is one of the most difficult topics in the Fundamentals course. What I’m about to explain isn’t as much a “drag n’ drop” issue as a “knowing where to click” issue. I still felt compelled to throw in this example as it’s a mistake almost everybody makes when they’re getting started with Tableau.

Let’s take a look at this example. Month of Order Date is green (Continuous), thus, sales is shown in chronological order over a time-series.

 

 

When you want to convert Order Date to a Discrete date, what you should do is click on the dropdown menu and switch it in the middle pane:

RIGHT:

 

 

Result:

 

 

Instead, what often happens is that people switch it down the bottom pane:

WRONG:

 

 

….producing this result:

 

 

Notice that even though your Order Date field on Columns switched to blue (Discrete), the behaviour itself is still Continuous. The year colour-coding is still in a time-series rather than piled on top of one another “in buckets” over Jan – Dec.

So what has changed? Look at the labels (x-axis) now treated as ‘discrete’ and compare that to the first screenshot.

There are times when you do need to employ the ‘bottom-pane’ Discrete – Continuous “toggle-r” (link to a real-life scenario below), but let’s keep it simple and forget the bottom pane exists for now.

Article link -> http://tridant.com.au/blog/replicating-google-analytics-using-tableau-flexible-yoy-rolling-period-comparisons-and-trending/

I hope that was helpful!

Share on ...

Ana Yin

Senior Consultant

ana.yin@tridant.com.au

1300 737 141

View more articles by Ana Yin

Leave a comment