Hey, all,
Here’s an interesting new concept from JMP’s Xan Gregg: Packed Bar Charts. See his blog post here: https://community.jmp.com/t5/JMP-Blog/Introducing-packed-bars-a-new-chart-form/ba-p/39972
These charts are a good approach to dealing with a lot of categories in a data set, especially for data where many categories compose a small fraction of the total. We’ve all seen pie charts where there are two big wedges and a dozen tiny wedges. This improves on that kind of visualization by focusing on the biggest categories, but still supplying information on the background categories for context.
A script is attached to Xan’s post, if you’d like to start playing with this chart paradigm. Feel free to ask me if you have any questions!
Here’s Xan’s explanation of the concept:
----
How do you visualize skewed data over hundreds of categories? I’ve been experimenting with a new chart form that could help with this. I'm calling it “packed bars.” It's a synthesis of an ordered bar chart and a treemap. Here’s what a packed bar chart looks like with recent S&P 500 market capitalization data.
The blue “primary” bars and the axis form a regular ordered bar chart showing the 12 largest companies by market capitalization. Each gray “secondary” bar represents one of the other companies. Some of the secondary bars can be labeled if space permits, but their main purpose is to provide context for the primary bars. The idea is to support the Focus+Context data visualization principle.
We can get a rough sense of how many secondary categories there are, what their total size is, and even how the sizes are distributed. And since they are on the same scale as the primary bars, we can even estimate some of the values from the length they span on the axis. In a dynamic display, category labels and values for secondary categories can be accessed via hover labels.
