News and Notes from PolicyViz - Issue #25

Hi all,

I think I can see the end of the year from here! Yep, the US Thanksgiving break is upon us, which means we are in the stretch run to the end of the year. I'm hoping to get a bit of a break during December to catch up on some writing, podcasting, exercising (what!? what's that?!), and some stuff around the house. In the meantime, check out a draft blog post below and some new things I'm reading.

Oh, and starting in 2023, I'm going to try to add a "job openings" section to this newsletter, so if you're looking to hire anyone, please let me know and I'll get it in the newsletter.

Happy Thanksgiving,

Jon

DRAFT POST: Showing Shares: Area Charts vs. Line Charts

Take a look at this stacked area chart from the Federal Reserve Economic Data (FRED) webpage at the Federal Reserve Bank of St. Louis. It shows the distribution of wealth in the United States between late-1989 and mid-2021 across six separate household income percentile ranges (0-20, 20-40, 40-60, 60-80, 80-99, and 99-100).

One thing that is clear in this view is the growth in wealth inequality in the United States, especially the share of wealth held by the top 1% of households. At the end of 1989, the share of total wealth held by the top 1% stood at 17.2%, ultimately rising to 26.9% by the middle of 2021. The share of wealth held by the bottom fifth of the distribution stayed roughly the same over the entire period, ranging between 2.2% and 4.1%.

There are a couple of things we can’t see in this stacked area chart. First, it’s hard to see the relative shares of each group. For example, how much greater is the share of wealth held by the top 1% than other groups? We can eyeball it a bit—in 2021, the share is clearly larger than each of the other quintiles other than the 60-80% group. Second, because of the line-width illusion, in which we tend to assess the distance between curves at the closest point rather than the vertical distance, it’s difficult to accurately assess the trends of each share over time.

Maybe a line chart a better choice? Honestly, I’m not sure. On the one hand, you can now see the wealth share of the top 1% is about the same as the share of wealth held by the 60-80th percentiles until about 2003, when the two lines separate. You can also more clearly see changes and the gaps between the shares over time. On the other hand, we don’t really get that part-to-whole perspective that the stacked area chart affords us.

I did wonder whether having the vertical axis go to 100% was necessary here, so here’s another approach with that vertical axis going to 50%. Maybe going to 100% is useful to imply (along with a good title and maybe some annotation) that these lines sum to 100% in each year? We obviously get a more detailed view of the data in this more restricted vertical range but again don’t immediately see the part-to-whole relationship that we get from the stacked area chart.

But perhaps that tradeoff is worth making? As with many aspects of data visualization, it really does depend. It depends on the patterns you want to highlight and the arguments you want to make. The stacked area chart helps focus attention on the part-to-whole relationship while the line chart focuses attention on the patterns of the individual series.

What do you think? Is either one a clear cut winner for you?

Please let me know if you have thoughts on this blog post. I post drafts here and then put them on my site about a week later.

PolicyViz Shop Sale!

Get 25% EVERYTHING in the PolicyViz shop all this week! Shirts, books, games, ebooks, and much, much more. Use the code TDAY2022 at checkout.

PolicyViz Podcast with Max Kuhn

Max Kuhn is a software engineer at RStudio. He is currently working on improving R's modeling capabilities and maintains about 30 packages, including caret. He was a Senior Director of Nonclinical Statistics at Pfizer Global R&D in Connecticut. He was applying models in the pharmaceutical and diagnostic industries for over 18 years. Max has a Ph.D. in Biostatistics. His latest book, Tidy Models with R, co-authored with Julia Silge, was published earlier this year.

Sponsor: This week's podcast is brought to you by Partnerhero. To waive set up fees, go to http://partnerhero.com/policyviz and mention “PolicyViz” during onboarding!

What I'm Reading & Watching

Books

• Stuck? Diagrams Help. by Abby Covert

• Index, A History of the: A Bookish Adventure from Medieval Manuscripts to the Digital Age by Dennis Duncan

• White Rage: The Unspoken Truth of Our Racial Divide by Carol Anderson

Articles

• QuantCrit: education, policy, ‘Big Data’ and principles for a critical race theory of statistics by Gillborn et al.

• Considerations for transgender population health research based on US national surveys by Lett and Everhart

TV/Movies

• The Crown on Netflix

Note: As an Amazon Associate I earn from qualifying purchases.

Check out the PolicyViz YouTube channel!

A few new videos recently, including an Excel tutorial and data visualization critique, but there is a lot of content there that you can use to help improve your data visualizations! If you're new to data visualization, check out the One Chart at a Time playlist, which contains more than 50 short videos on different graphs, charts, and diagrams.

I'm also considering creating a set of videos to the basics of creating data visualizations in Excel. I'd demonstrate the core, built-in graphs like bars, areas, treemaps, and more. Just targeting the introductory user to help them be better at Excel. Let me know what you think!