10. More Charts in Tableau

Overview

In this lesson, we'll explore some of the remaining chart options which we haven't utilized in this course so far.

To explore more Kubicle data literacy subjects, please refer to our full library.

Summary

Line charts

- Used primarily with time-series data
- Can be either discrete or continuous
- Continuous line charts need to be watched carefully for zero values

Area charts

- Also used primarily with time-series data
- Can be either discrete or continuous
- Continuous area charts need to be watched carefully for zero values
- Discrete area charts can underscore the value of isolated data points
- Consequently, these charts are best used when we know that no zero values exist

Packed bubbles chart

- Excellent visualization for showing large number of data points
- Particularly useful when identifying a client's exposure to the larger values in a dataset
- Can also be color-coded effectively using dimensions

Pie-charts

- Typically inferior to packed bubble charts, especially for larger datasets
- Sizing in Tableau doesn't work well and pie-charts can appear small on screen
- Try to avoid this visualization if possible

100% stacked column chart

- Another alternative to the pie-chart, works well for visualizing changes over time
- Not as effective as the packed bubble chart for larger datasets
- However, better than the packed bubble chart for time-series data

Transcript

In the previous lessons of this course, we've covered almost all the different visualization options available in Tableau. There are a couple other options left in the Show Me dropdown that I'm going to cover in the remaining two lessons. Let's start with two charts that we covered in the first course which are popular for visualizing time series data: Line Charts and Area Charts. When using these charts, the date field is normally in the column shelf and the remaining measure, say revenue is in the rows shelf. Tableau has a nice way of enabling you to separate your line charts into discrete time periods such as quarter or month. In this example, I have my value showing by week. If I'd like to split this chart into quarter or month, simply take Date field and drag it into Columns.

This will automatically set the time period to year. I'd hit the Plus sign so it'll give me quarter and then hit it again to give me month and this now separates our line charts into these time periods. This type of line chart is called discreet. I can also show a continuous line chart where none of these segments are created.

However with continuous line charts, we need to be very careful particularly when we have zero values in our line chart.

Let's take an example where we specify a number of different clients removing the existing selection and taking a couple of different companies like so.

One thing to watch with continuous line charts are their treatment of zero values. In this example, I've selected five specific clients and shown their fees plotted by each division. At first glance, it appears as if advisory fees have shot up from August to October. However this is not the case. If I hover over the data points at the end of each line, I'll see that I have one data point on the 25th of October and another data point on the 9th of August. But no data points unfortunately in between. So this chart is very misleading. It appears as if fees from this division are rising steadily but in fact there are only two specific data points. This becomes clear when I flip this chart type to a bar chart where we can very little advisory revenue on the 9th of August then a lot on the 25th of October and nothing in between.

As a consequence, when you're using continuous line charts, you need to be very careful that no zero values exist because the propensity to misinterpret a chart is very real. Area charts also struggle to visualize data sets such as this. When I flip this chart to an area chart, be it discreet or continuous, visualizations of single data points, for example advisory on the 25th of October, appear as a single line.

And unfortunately, this makes the data appear almost irrelevant when we know from the bar chart that this data point is most certainly not irrelevant. So if you have zero values with time series data, you may need to use a bar chart rather than line charts or area charts to make sure that your visualization is not misleading.

Another visualization we used in the first course is the packed bubbles chart. This works really well when we need to visualize a lot of data points and understand how big a contribution the largest values contribute. This can also be color coded to create some additional insights. Pie charts are often used to visualize the same type of data but to my mind, when used are much harder to interpret than a packed bubbles chart. If we want to increase the size of the pie chart, we need to change the fit from Standard to Entire View.

When we moved to Entire View, it will increase the size of our pie chart to fill the screen. As you can see in Tableau, it doesn't actually give the pie chart a lot of space to build compared to the packed bubbles chart. What's more, when we have so many data points, it's incredibly difficult to interpret this slide and any detail.

If you are trying to pick between these two different chart types, I would normally go with the packed bubble more often than not unless the number of data points is very low. Another alternative to the pie chart and the packed bubbles chart is the 100% stacked column chart which we built in an earlier lesson. This works very well when we want to visualize changes over time and we don't have a lot of data points to plot. To create the percentage of total value metric in this type of chart, I'm going to add a label to show you how to do so. And that label will be Sum of Value. So I'll add Sum of Value onto Label. Unfortunately, this doesn't give me the label that I want. I want the percentage of the total fees by division for that particular month. So to do this, I'll hit the dropdown, I'll go to Quick Table Calculation and Percent of Total. This gives me the percentage of the total for the entire sheet whereas I want the percentage of the total for the entire month. I'll hit the dropdown again. This time I'll go to Compute Using and select Table Down.

This now gives me exactly what I want. And to create the 100% stacked column chart, I simply repeat that from a normal bar chart up in my rows shelf. Simply going to the Quick Table Calculation and selecting Percent of Total and compute using Table Down. As you can see from this chart, it's now very easy to understand how fees by division are changing over time. For example Audit has moved considerably as a percentage of the total from August to November. Tax has stayed pretty much consistent and Advisory has dropped substantially. In the next lesson, we'll wrap up our look at some of the other charts in the Show Me dropdown including Gantt charts and bullet graphs.

>
Tableau Essentials
Creating Visualizations in Tableau

Contents

My Notes

You can take notes as you view lessons.

Sign in or start a free trial to avail of this feature.

Free Trial

Download our training resources while you learn.

Sign in or start a free trial to avail of this feature.

Free Trial