From CS294-10 Visualization Fa08

Jump to: navigation, search


Good Visualization


Photo by Brian DeRenzi.


This is a bar chart of the number of patients with various diseases in a small health clinic in Uganda. It's simple but effective in showing the severity of malaria in the area. The data is the count of patients grouped by disease and ordered by frequency. For a hand-made visualization, I think it's really cool. A pie chart may be good for this, but that's harder to make by hand.


The data for the visualization is the number of patients diagnosed with a particular disease. There is no indication for what time period, so I would guess it's the total number cases in the clinic recent history. The Y dimension is a count: total number of patients. The scale is linear. The X dimension is categorical: diseases including malaria, acute respiratory infection (ARI), pneumonia, eye infection, diarrhea, anaemia, skin infection, ?, intestinal worms and ear infection. This dimension is ordered by prevalence of the disease. Some of these categories are themselves aggregates in a disease hierarchy; if the visualization were digital, perhaps a nice drill-down would be a breakdown of ARI patients by tuberculosis and other respiratory infections. The categorical granularity is not uniform: malaria is a single disease, whereas many other categories cover multiple diseases. While this is not obvious, it serves cohesion of the visualization: even lower counts of non-malaria diseases would not necessary underscore the extremity the malaria problem. The two circles and "starburst" decorating the center of the graph does not provide additional information.

Okay Visualization


Photo by Brian DeRenzi.


This is a table found in the same health clinic showing how many vaccines they received and how many vaccinations given. Each month the staff enters the units of received vaccines and total vaccines given. It seems what they really want to do is to allow the staff to track the difference between vaccine intake and usage. The don't have a computer in the clinic, but still want the new data visualized for the staff -- so they they use a table form.


This visualization is two tables. In both tables, the columns are months of 2008, and ordered as one would expect from Jan to Dec. The rows are types of vaccines supplies received and vaccinations given to patients. The data is count total of ambiguous units of received and dispensed vaccines. Each Y-axis category in the Vaccines Received table has two sub-rows, but the 2nd sub-row is used only to track a 2nd type of diluent.

There are several pieces of missing information. The units of vaccine intake and usage isn't specified. Certainly they are not the same units. As well, there is no accounting of diluent usage: some amount is received for measles and scg sub-categories, but none specified for usage - another units problem.

There is an incomplete table entitled "TARGET". As in the first visualization, there is a decoration in the otherwise empty space. This time it's an encouragement to "Stop the outbreak". This serves as a reminder to those who use the visualization.



In my reconstruction, I put together the two tables that had almost same columns and rows. This way users can more readily compare input and output. The units are still unknown, so I took a best guess and put the units in the a Key at the top of the page. I used colors to indicate what each sub row means under a particular vaccine.

For a visualization tracking inventory, there should be an indication of the total stock, so I added a space for that in green.

Vaccines requiring diluents are now closer and better associated with the vaccine's data.

As well, I removed the Target table, and added a column in which the users can track target totals.

Bonus Visualization

Image:a1_kuangchen3.jpg New York Times

I have been obsessed with this visualization for a while. Let's just focus on the top bar:

  1. This is great way to visualize the poll numbers in context of the electoral college.
  2. The black bar in the middle, and its position with respect to the yellow band of "tossup" voters
  3. The bar also serves as a color key for the map visualization below

[add comment]
Personal tools