If only Vizhealth were fully functional...

First of all, this report on the Visualizing Health program ran by the University of Michigan is incomplete because much of the website is currently nonfunctional. 

To start, the program seems to be for a very good cause: the website serves as a guide for presenting medical information to people, particularly through imagery. They provide visualizations, which have been shown to be effective via scientific studies, under Creative Commons Attributions, meaning that you only have to credit them. 

The gallery is one of the parts currently down, but from this informational comic visible on their website, the visualizations that they provide seem to be quite helpful. There seems to be many different charts each adapted for a particular facet of medical reporting. For example, they offer a basic risk calculator for cardiovascular disease (based on a real scientific formula) and the final display makes great use of size and color-coding. They also provide a numeral and a pie chart, making it comprehensible to those with low and high numeracy skills, as discussed in chapter 2 of "Effective Data Visualization."

Moving on, I believe the most educational part for us is the icon array generator that they include.

Here is an image displaying all of the different configurations that they offer, and I think it is worth exploring some of them in particular. 

  I wonder if there is any psychological difference between 100 versus 1000. With these two charts the percentages are the same, for your information. 

To make use of this chart some more, the options for male, female, and gender-neutral is really nice and personalized. 

I am not sure if this feeling is justified by anything, but I much prefer the person icons rather than the offered solid block icons. Perhaps it is the unfulfilled need for visual health data to get personal that bothers me. 

What I find the most fascinating, however, is the ability to randomize the positions of the colored icons. The axis label is quite useless now lol. Anyway, I feel like this may be a better way than just clumping the sick people together at the bottom (or the top as you can change it to that); I feel that the clumping causes you to mentally divide the array into two sides and it's harder to compare and intuit the proportions. With the randomization, you see how the sick are distributed within the entirety of the population and there's also an effect of increasing perceived urgency because the randomness signals that it could happen to anyone, including you. However, this is all speculation and I am not sure if the evidence actually supports my ideas. Still, I think this might be the most important setting to consider for this case of icon array generation. 

However, ultimately, the website in its current state is not of much use to data visualization designers as the gallery and many other links do not work. The only educational part, I feel, was our experimentation with the icon array generator. How tragic...

Edit 9/23/25, Mrs. Wendt, were we supposed to actually make a real storytelling visual with the icon array because all I understood was experimenting with the generator. If you'll give me the chance, I could make one, and if not it's okay.