Bar Charts - A User's Guide

How To Use Bar Charts

Bar charts are some of the most common and useful data visualizations. They can be used to express a wide range of data in a wide range of relationships, and many BI dashboards have at least one or two of them. 

With the rise of self-service BI and user-focused dashboarding tools in software like Domo or Tableau, more people than ever before are interacting with data and building their own dashboards. However, many of these users don't have very much data experience. When they build dashboards, they lean towards basic visualizations like bar charts to express their data. 

This isn't always a bad approach. Bar charts are fairly customizable, and those that are intimidated by the massive list of visualizations in a modern BI tool are better off using a basic visualization like a bar chart than trying to do something too fancy. However, this means that many users default to bar charts for showing their data, even when something else might be more effective. 

Bar charts can be a powerful tool, but just like every other data visualization, they have their strengths and their weaknesses. With a better understanding of bar charts, when to use them, and when not to use them, dashboard designers can build more effective dashboards and help their users to understand data better. 

What is a bar chart for? 

Bar charts show a comparison between the same variables in a data set across different categories of data. In human terms, you're comparing a statistic between different groups, cohorts, regions, countries, and so on. 

Imagine that a manager wanted to compare sales figures across different regions. They want to compare the same statistic (sales) using the same metric (income) across different categories of data (departments). This is the optimal use case for a bar chart. 

Bar charts can also compare some variables over time. For example, a sales manager could see how total sales compare to each other week over week. In this way, a bar chart acts very similar to a line chart, which also tracks a variable over time. 

The key difference is the goal of the visualization. A bar chart is better at showing the raw comparison between two dates, since a user can visually compare the size of the bars. It's a simple shorthand to tell how much bigger or smaller one variable is compared to another. 

A line chart can often show the same data set, but the goal of what's being communicated is different. A line chart is better at showing the change week over week, and its severity. With a line, it's easier to see spikes, dips, and plateaus. 

That's not to say that a bar chart can't show this sort of information. However, in general, a bar chart is a better choice if you're worried about a data set's magnitude, and a line chart is better if you're worried about direction. 

Grouped bar charts

Dashboard builders can also use bar charts to compare multiple cohorts across the same category and compare that category over time. A grouped bar chart is similar to a regular bar chart, but all of the bars that represent the cohorts have been squished together and the X-axis is time. 

In this way, a grouped bar chart is like having an individual bar chart that compares cohorts for every time period. Many businesses use grouped bar charts for many different reasons. For example, a marketing executive might want to compare performance over different channels, and see how that changes week to week. That's an excellent use case for a grouped bar chart. 

Often, novice dashboard builders feel like they should use a line chart for this use case, with a line representing each cohort. In general, multi-line charts are much harder to read than a grouped bar chart. Unless it's important to specifically show the change over time, it's better to use a grouped bar chart. 

Read More