When Dollars & Sense Magazine hired us to create custom data graphics for its December 2015 cover story, our first task was choosing the right approach.
Author Dan Schneider had argued his main point quite clearly in the cover article, “The Worst Place in the U.S. to be Black is Wisconsin,” using supporting evidence from a broad survey of social-economic studies. So, before doing anything else, we needed to determine which class of data graphic would visually drive home the cover story’s thesis.
We resisted the inclination to create a visually extravagant graphic with datasets plotted from each source – knowing that it could overwhelm the main point with a sea of disparate statistics. Instead, we sought a design that would emphasize Wisconsin’s miserable standing compared to other states. This seemed like a job for a special class of data visualization: the slopegraph.
In our case, the key feature to display was not the actual value of each state across various statistical measures – it was simply the relative rank of each state to its peers. Therefore, unlike most slopegraphs which plot continuous data on one or two y-axes, we chose an ordinal ranking of states across all statistical measures with a uniform scale on the y-axis:
Then, to clearly show how poorly Wisconsin ranked, we emboldened the lines and type for the three states that came in worst place on any one of the measures (New York, Minnesota, and Wisconsin) and used a spare amount of color to highlight Wisconsin’s particularly sad place among its peers.
The trade-off with this design is a loss of information on each measure (what is Wisconsin’s African-American incarceration rate?) for a clear comparison of states across multiple measures.
For more on slopegraph variants and design considerations, check out Charlie Park’s fine essay on slopegraph typology.
December 9, 2015 update: My good friend and toughest critic, Micheal Tofias, takes exception to calling this a slopegraph:
I am pretty sure that the point of the slopegraph is that the slopes are meaningful. That’s why they are often used to connect the same variable measured at two or more time points (across some number of subjects). The slope is the rate of change.
In your work for Dollars & Sense, the slopes are not directly or easily interpretable. At best they represent something like consistency across the various rankings.
I am sure this kind of plot has a name, but I am not sure what it is.
A fair point. As noted, this graph plots the rank order of the states across various measures, not the actual values of the measure for each state. That, plus the fact that the measures are not comparable by time, certainly means that the lines convey much less information than most slopecharts. In fact, what the lines are actually doing here are simply serving as extended labels that connect the rank values of each state across all measures.
I could call it a “bumpchart,” which Parker considers a particular subset of slopecharts. But again, it does lack the temporal component. So, maybe I’ll just call it “custom” and leave it at that.
In need of a custom data visualization? Let us help. Get in touch.