Tableau: Visualize Continuous Variables

This document is the Tableau implementation of the more general, conceptual discussion regarding data visualizations of continuous distributions.

Histogram

Tableau has no histogram function per se. There is no pre-defined histogram plot, just a bar chart. Instead, follow these steps.

1. Identify the continuous variable of interest in the list of variables on the left-side of the worksheet.
2. Create a new categorical variable version of the continuous variable, a binned version of the original variable.
   [Right-click on the variable name and select Create and then Bins...]
3. Create a bar chart of the new categorical variable.
   [Drag the (Count) variable to the empty shelf.]
4. Make the bars wider on the bar chart until they become adjacent to each other with no gaps or very small gaps that act like a border between adjacent bars.
   [Select all the bars, go to the Size mark, and drag the slider.]

The link to the video of this process follows.

Video: Histogram [4:31]

Box Plot

The Tableau default box plot is the box plot with the plotted points. There is no jitter option. Create with the following steps.

1. Drag the variable name for the continuous variable of interest to the Rows or Columns shelf depending on the desired orientation of the box plot.

Be aware of the following potential “gotcha”.

Tableau automatic aggregation

Tableau often aggregates variables by default, identified such as by SUM(name). Unaggregate from the Analysis menu, then deselect Aggregate Measures.

Figure 1 show the result of the automatic aggregation after dragging the Salary variable to the Columns shelf. When you unaggregate from the Analysis menu, the SUM label disappears, leaving only the variable name.

Figure 1: Automatic aggregation, which defaults to sums, as shwon by the cumulative sums of the salaries along the axis.

Tableau also remembers how you treat a variable, so once unaggregated it may be remain so with further analysis. So, depending on your recent history of working with Tableau, you may or may not need to unaggregate. For some reason, there is a way to change the aggregation by right-clicking on the variable name when in a shelf but not a way to turn off the aggregation without going to the main menu.

2. Click on the small Box Plot icon in the Show Me tab that displays the available plot types, shown in Figure 2.

Figure 2: Tableau box plot icon.

3. To create a Trellis plot, drag the categorical variable of interest to the opposite shelf that contains continuous variable upon which the box plot is based.

4. If desired, start the box plot at the minimum data value instead of 0.
   [Select the axis by clicking on it, right-click on the axis and select Edit Axis, then uncheck the box labeled include zero]

5. If desired, if the points overlap, add some transparency to the plotted points by selecting the Color mark, then slide the Opacity slider to achieve the desired transparency.

The link to the video of this process follows.

Video: Box Plot [3:44]