On this Advanced lesson on Tableau Box Plots, you will learn what Tableau box plots look like, how to make them, and how to interpret Tableau box plots.
How does a Tableau Box Plot look like?
A Tableau box plot looks like this:
It allows you to compare a range of values across several segments. Let’s say we wanted to see the breakout of ages for our employees. We could use a histogram or bar chart to show how many people fall into each of the age buckets. If we wanted to stratify salary ranges, we could take the same approach. But, what if we want to see salary ranges per age range? That becomes a much harder problem to visualize.
Tableau box plots are a simple way of accomplishing that. They show ranges of data, or distributions, across one or multiple segments. In the chart above, we can see the distribution of salaries of people in their 20’s (the first column in the chart). We can see that the average salary is just shy of $40,000, but that we have some outliers at $100,000 and $15,000. The $100k per year 22 year old might be a data scientist, fresh out of school with a dual major in statistics and business. The 22 year old making $15,000 probably didn’t go to college and is working more of a minimum wage type job.
How to Interpret Tableau Box Plots
The line in the middle of the shaded Tableau box, or the dividing point between the two colors, is the median or midpoint of all the data values in the range.
The shaded area on each set of dots contains the middle 50% of all the data. This shaded area is known as the Interquartile Range. The bottom 25% of the data is below the shaded Tableau box. The top 25% is above the shaded Tableau box.
The whiskers typically represent 1.5 times the min or max of the shaded Tableau box, or interquartile range (IQR). So, the bottom whisker is 1.5x the min of the IQR, and the top whisker is 1.5x the max of the IQR.
The points at the very end represent outliers, if there are any. These are points that don’t follow the rest of the distribution. They are definitely worth investigating, and can often be the most useful pieces of data in your set. They indicate some anomaly in the data, like a data error, or they indicate examples where the normal pattern breaks down for a good reason, and understanding why can lead to major new insights. If you decide they are skewing your data too much, you can exclude them to focus on the otherwise normal patterns.
How to Make a Box Plot in Tableau
- Drag Region to Columns and Profit to Rows. This will build us a bar chart by default.
- Drag Customer Segment onto Columns, to the right of Region.
- Click on Show Me and choose the Tableau Box Plot option.
Tableau will automatically build you a Tableau box plot. However, it may not be organizing the pills in the order you want. In our case, we want Region on Columns and Customer Segment on Detail.
- Click on the Analysis menu at the top.
- Click on “Cycle Fields” towards the bottom of the menu. This should create the right Tableau box plot structure, with Regions as our headers running along the bottom, and each circle representing a Customer Segment.
- Now, move the Customer Segment pill from Detail to the Color shelf, by dragging it in.
- Set the view to Fit Width from the drop down menu above the chart.
- Title your chart “Profit Distribution by Region and Customer Segment.”
- Lastly, let’s size up our dots just a bit, by clicking on the Size shelf and moving the slider up somewhat.
We can now very quickly see that in most cases, Corporate customers are the outliers. They are our most profitable. Conversely, Consumers are our least profitable. Except in the Mid-Atlantic region, where Consumers are our top. This might be worth further investigation.