Blog

BLOG

A focus on visualization: Bar graphs

PUBLISHED ON 5 MAY, 2014

Nowadays, there are so many ways in which data can be visualized to allow you to easily comprehend it and take insights from it. Everyday new visualization platforms, ideas and graphs are appearing. Although these are often visually appealing and can create great value, it is vital to ensure you are selecting a graph that best represents your data.

Sweetspot would like to recommend a few tips to make sure you use the best visualization for your insight delivery need by exploring a number of graph types.

To kick-off our new series off we’d like to go back to basics with one of the most common graphs: The bar graph.

A familiar design to most, the bar graph is used to represent categorical data along a nominal or ordinal scale. Bar graphs are easily interpreted by comparing the length of the bars, which allows viewers to spot relationships quickly as opposed to the effort associated with comparing areas (something humans are not very good at).

Additionally, sorting the bar graph by values (in ascending or descending order) will make comparisons even easier. As quick tips:

When to use a bar graph	When NOT to use a bar graph
Bar graphs are best used to represent: Comparisons of data happening at a static point in time. Distributions of data when time is your only dimension and you want to stress its specific values.	Bar graphs are generally not our best option when: Distributions of data over time with a focus on the trend. We want to measure the contribution of each value to the whole. Time is not your only dimension.

Types of Bar Graphs:

Grouped bar graphs: are used to compare the same variables across multiple categories. If you’d like to see which pages of your website are most visited by device-specific audiences, this the graph for you. First decide which category you would like to look at overall, and then decide what factors are important to you. In this example you can see how your conversion rate breaks down between multiple sources.

Stacked bar graphs: present information in a sequence, allowing the reader to not only compare, but also view the data breakdown of each category. For example, If you would like to see how your social media outlets are engaging with the public quarterly:

 

Important things to consider when creating your bar graphs:

 

• Scale and intervals: Set a scale that best represents your data based on your lowest to highest variable. Remember you don’t always have to start at zero, although it can avoid confusion. Also, be careful to adjust your intervals so that your bars are easy to see.

We split our main axis in intervals of 50 in the graph above, but the data is bunched together. Now look at the effect of “zooming in” and using intervals of 25:

Do you see the difference? It becomes much easier to spot the distribution of values.

Keep in mind, you can also use two scales at the same time to correlate data with very different number ranges.

For example, in the graph above you can see two scales used to measure percentage, in comparison to visits/visitors.

• Dimension: 3D, or not to be? Although it might be tempting to try to spruce up your bar graphs with visual effects such as 3D bars, approach this option with caution, stay away from 3D! 3D graphs will either make your data pop, or leave your reader stumped.

 

• Color: It’s easy to get carried away with so many color choices, but before you choose your color palette ask yourself what message you are trying to get across. Color should communicate, define, and highlight important data.

 

Colors should allow you to distinguish easily between different indicators. Notice how in the example above the all black graph forces the reader to waste time working out which bar is which, making it very ineffective. Remember: The right colors make data easier to read!

• Background: As well as color choices, background choices can sometimes get the best of us.

The graph above is a perfect example of why less is sometimes better.

In conclusion:

It’s our hope that you optimize your data delivery experience by choosing the best visualization to fit your needs. Selecting the right type of chart will help your stakeholders (“data consumers”) to easily and quickly understand what is happening, and this increases the likelihood that they will act on it!

What else do you take into consideration when building the perfect bar graph for your reporting or insight delivery needs?