Bar in Bar Chart in Tableau #BestPractices #VizTips
Tableau can create some very special comparison charts. One such chart that has gained wide recognition in the recent past is 'Bar in Bar' chart.
The concept of Bar in Bar chart is very simple. We focus on typically 2 measures (or more*) with one measure which acts as the main or actual measure in consideration and a threshold measure against which the main measure is compared.
To follow best practices (*Not rules or regulations, only practices), it is good to have the threshold on the background and the actual measure on the foreground.
To demonstrate this, a small sample data is selected. There are two measures. Actual is the main measure (will be used in front) and Forecast is the threshold measure (will be used in the back). The Dimension Name is used for slicing and dicing.
This is the most widely used method for creating a Bar in Bar Chart in Tableau.
This method uses 2 special fields called Measure Names and Measure Values.
Just a quick recap on Measure Names and Measure Values:
Measure Names and Measure Values are auto generated fields within the Tableau UI. (They are Italicized and not Sorted in alphabetical order when compared to the data source fields.)
1 Measure Value field acts as a single shelf (or bucket) for all the measures in the data pane including Number of Records.
1 Measure Value field also acts as a placeholder for incoming measures.
Measure Values field is under Measures pane. It is always Green in color representing a Continuous field. Measure Values itself will not get aggregated, instead stores the aggregation for all measures. Measure Values field in most cases represents a scale or an axis.
The number of measures that can be displayed on the visualization is controlled by Measure Names.
Measure Names is always Blue in color representing a Discrete field.
For many comparison charts, Measure Names will be on the Filters shelf and/or Color card and/or Size card.
Measure Names represents names for all measures and color buckets
Measure Names and Measure Values are Name-Value pair that go hand in hand for most comparison charts.
All these features of MN and MV will be exploited in this first method.
There are different ways. But here is the step-by-step approach for one such way.
1) Use Measure Names on Filters shelf
2) Select the 2 measures (Actual and Forecast) required for this visualization. (Looking for comparable measures only, and not metrics of different scales. E.g. Revenue in Millions and Percentage of Returns as a % cannot be used for this example.)
3) Use Measure Values on Rows (or Columns). For Vertical chart, use Rows; for a horizontal chart, use Columns. (At this point, a new Measure Value shelf is created, both the measures standing together, one behind another). To understand the order of operations, the measure on the top will be in front. It can be rearranged later.
4) Use the Name field on Columns (or Rows) for slicing the aggregation into multiple pieces
5) Use a copy (one more) of Measure Names on Color card. (* Creates a Stacked Bar)
6) Use a copy (one more) of Measure Names on Size card. (This make one bar thicker than the other)
7) Now for the last step. From the Analysis menu on top, “SWITCH OFF” the stack.
This method will act more like a hack or a trick. Here is the step-by-step approach:
1) Use the Forecast measure with SUM aggregation on Rows (or Columns)
2) Use the Actual measure with SUM aggregation also on Rows (or Columns).
To understand the order of operations, the measure on the left will be on the back. It can be rearranged later.
At this point, the Marks shelf is split into 3 components, one for each measure and one for All.
3) Use the Name field on Columns (or Rows) for slicing the aggregation into multiple pieces.
4) Use the second measure to create Dual Axis
5) Use the “All” segment of Marks shelf and convert to Bar chart
6) Use the “Actual” segment of Marks shelf and move the slider to the left (on Size card) to make it look like a Bar in Bar chart
7) Now for the last step. Synchronize the second axis.
This method also acts as a hack or trick. This method needs a little work on the look and feel of the viz.
Here is the step by step approach.
1) Use the Actual measure with SUM aggregation on Rows (or Columns) .
2) Use the Name field on Columns (or Rows) for slicing the aggregation into multiple pieces.
3) Use the Forecast measure as SUM aggregation on Detail card (This step is a must. Without this step, the future steps do not fall in place)
4) Use a Reference Line (or Average Line) from the Analytics Pane with the following features:
a) Per Cell
b) Without Labels
c) None for the Line under formatting
d) Fill below color (Pick the appropriate color of choice)
(The Reference Line has to be created for the Forecast measure. This measure now goes on the background. In this case, Sum, Avg, Min, Max, Median all work the same.)
5) Reduce the size of the bar for the Actual measure using the Size card slider (Marks shelf).
6) From the Format menu, we need to focus on Grid Lines:
a) None for Rows
b) Thick white grid lines for Columns
7) If needed, a small annotation or caption legend can be provided on the viz or for the sheet
Here you go. I hope you liked it. Let me know your feedback.