How to fix a multi-colored stacked bar chart?

A stacked bar (or column) chart is basically a pie chart unrolled to make a stick.

The challenge with stacked bar charts is they are usually not very effective. When we use stacked bar charts it appears as if we are attempting to answer too many questions with a single chart.

In this article, let’s look at one way to fix this problem.

Uses and limitations of stacked bar (or column) charts

Stacked bar charts are good when we want to compare totals across different categories plus within a given category, we want some understanding of the individual sub components.

What are its limitations?

Stacked bar charts are not good to compare the individual sub components across the categories. This is because beyond the first category (first and last in 100% stacked bars), there is no consistent baseline.

This is the main drawback of stacked bar charts. If there is anything of interest happening further up the stack we cannot see it because it is stacked on top of things that are also changing.

How to fix a multi-colored stacked bar chart?

stacked-bar

The problem with using so many colors in a chart is it introduces a sense of complexity. People perceive more aesthetic designs as easier to use. They tend to be more engaged with a chart if it appears pretty.

Of course, we could use a different chart type – but let’s assume our audience demands to see all the data and we are short on space.

Here are some ideas to make a stacked bar like the one above easier to interpret.

stacked-bar-makeover

Identify your focus. Stacked bar charts can be used to compare the totals as well as understand the contribution of sub components. But they cannot do both with ease. It is important to identify what your focus with the chart is and arrange the chart accordingly.

In this example, my focus is the total cost, hence the chart is arranged in descending order of total costs. As you can see it is hard to compare the cost of long drinks for instance. Thats a compromise we need to make with stacked bar charts.

Limit the number of sub components in the stack. Stacked bar charts work best with 2 or 3 sub components. If there are too many sub components in your chart, try grouping tinier parts into “others”. In this example, since there are only 4 sub components, I decided to show them all.

Put the piece of most interest in the baseline so it’s easier to compare across categories.

Use color with discretion. Highlight the most important piece with color and subdue the rest using greys.

Can you think of any more tips to add on? Let me know in the comments below.

You Might Also Like

  • #SWDChallenge: a makeover

    The #SWDChallenge this month is a makeover project. The challenge instructions come with a wonderful article by Elizabeth Ricks of SWD on the art of undertaking chart makeovers. I found the article very insightful and I took a lot of notes. In a nutshell, here are a few steps to get started with doing “good…

  • Should you use 3D charts?

    The short answer is NO. With all the magical tools available at our disposal, it is very easy to create awesomely complex looking charts and graphs that add no value. 3D charts are one of them. Edward Tufte called this conventional graphical paraphernalia that does not tell the viewer anything new. Furthermore, they come in…

  • Consistency in charts

    In this post let’s talk about the importance of consistency when telling a story with charts. So what is consistency? Consistency in story telling comes from colors, type of charts, font and size of messaging, alignment of various chart objects, to name a few. Consistency is everything that makes a chart look neat and easily readable. Let’s look…

  • Comparing change through time

    A clustered bar chart is one of the most commonly used representations to compare change through time. A clustered bar chart, though a good option to compare values, does not represent the change through time effectively. Here is some data on which technology products consumers intend to buy in 2016 and how it compares against…

  • Preattentive Attributes in Visualization – An Example

    Preattentive processing occurs without our consciousness at an extremely high speed. It is tuned to detect a specific set of visual attributes called Preattentive Attributes. If you want to understand more about how we process a visual, please refer to this post on Visual Perception. Preattentive attributes determine what information catches our attention. This is important in…

  • Sketching survey data

    The creative part of my brain has been asleep for a very long time 🙁 If you are like me, one way to get your creative juices flowing is to sketch on paper. This is a post with sketches on how to visualize survey data.     There are so many ways to show survey data….

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.