Posted by I work for University of Iowa Healthcare (all opinions always my own), and I love that I get to work with all sorts of departments from across the institution. Generally, my work is focused on operational analysis. I enjoy this, as it’s a strong suit of mine, but it does mean that I rarely have opportunities to map data.
I recently started working on a high-level project for senior leadership, with the goal of telling our story to a broader audience. A big part of that story is that although we are physically located in Iowa City, we serve the entire state of Iowa. This is because we’re the only Tertiary Care center in the state. To put it simply, whenever someone in Iowa gets too sick for their local hospital, they get transferred here. We really are a resource for the whole state of Iowa (and beyond).
Getting to this blog entry’s point, what this means is… Maps!
Maps!
The reason I even started writing this entry was because of this graph I created accidentally in my data flow:
I love how this emphasizes our regional (and national) impact. Also, as a design nerd, I love that it includes an image of Iowa via white space. I think this visualization may be an important part of the narrative, but we’ll see whether it makes the cut, as we’re still early in the project.
To skip ahead to the punch line, here is where I eventually ended up:
I don’t normally love labels, and when I do use them, I tend to use one or two. In this case, however, the visualization is intended for PowerPoint decks, meaning you can’t rely on hover texts.
A technical aside…
One technical point worth mentioning is how I addressed the color scale. Though we serve the entire state, we do have more visits from Johnson County (where we’re located). If I just used a straight color scale, it wouldn’t be all that informative:
To get around this, I set up a table calculation that takes the 95th percentile of the volumes for all of the counties, and uses that as a ceiling for the color scale:
There is an argument to be made that this is misleading, but I think the narrative intent behind the visualization (e.g. the story being told) warrants this design choice.
Short entry this time, but I hope you found it useful, interesting, or ideally both! Until next time…
Jeff
Infectious Data Viz 