A1-NateAgrin
From CS294-10 Visualization Fa07
The Good: 'Moving Target' National Geographic July 2007, pg 55
Moving Target diagrams the life cycle of the malaria parasite. This visualization is successful because it communicates many aspects of the stages of a malaria infection and where treatments might be successful. The visualization also demonstrates how one might actually see the infection at certain stages of the parasite's life under a microscope, and at what life stage does a parasite infect new mosquitoes. By showing the microscopic view of the parasite the viewer is exposed to otherwise unseen information and gives them a definitive visual reference point a points in the life cycle. I appreciate not only the cyclical nature of the visualization, but also the balance the visualization provides. There is a clear and even distribution between the mosquito halves of the life cycle and the human halves, which tells the information tell a story: malaria is not simply a human disease, but follows a complex, biological pathway incorporating multiple stages and hosts.
The Bad: 'Most Glaciers Losing Ice' National Geographic June 2007 pg 60
This visualization attempts to show the rate of change in average thickness of glaciers from around the globe. It's a fairly straightforward visualization, which could have a more lasting impression, if not for a few issues. The x-axis (time in years) and y-axis (change in thickness in meters) of the graphs are displayed only near the top of the visualization, and the origins of each graph appear to be random. If the visual information were somehow more easily compared, by providing a common origin or by plotting the information in one set of x and y axes, the overall downward trend of the total graphs might be more emphasized. This, I believe, would help tell the story of the information more obviously. As it stands however, the data gets its point across, just not very clearly or with emphasis.
The Bad Redux:
What bothered me initially about the bad visualization I chose was that, even though it followed Tufte's principle of the small multiple, it was very difficult to compare the graphs since they shared no common Y-origin. I corrected that by laying out the graphs in a 4x4 grid. To me, this makes it much simpler to glance through the data and see obvious outliers. In trying to 'maximize data-ink' (for what that's worth) I've reduced both axes to simple tick marks. I think this does improve the aesthetic value of the image, but I'm not convinced that it makes it any easer for the viewer to read. Another positive result of this format, is that a mean value can be chosen for the maximum (I chose -30). Graphs with a smaller value than -30 run off their allocated space, but simply fall into the visual space of the next multiple, eliminating the need for excess axis points which would only be used by the minority of the data. Finally, I plotted each graph in descending proximity to the equator. This adds another implicit piece of information to the overall visualization. As the reader scans from left to right and top to bottom they start from very close to one of the poles, and end up close to the equator. Regarding the single positive (that is, larger than zero) data point, it does not bother me in this format that it expands above the columns of the other data. In fact, I think this highlights the fact that it is an outlier and invites inquiry.
