Multivariate Display

For me, this was the most difficult and challenging visual we've done all semester. I used my pie chart from the original simple graph I did earlier, which displayed data for the high school GPA of first year RWU students. I had a lot of trouble thinking of what "story" to tell with this information as I searched for two other sets of data to use in the multivariate. I finally decided to make the entire visual about first year RWU students. I used data about graduating class rank as well as student location (in-state or out of state).

It was really difficult trying to make a multivariate graph with a pie chart. I couldn't think of any way to display all of this data. I wasn't showing information over time, so that eliminated a lot of possibilities for using different graphs. The data was very different from each other and didn't have the same variables, so it was really confusing trying to tie everything together. I'm still not even sure if the visuals I made even work as a multivariate display.

Even though I was unsure, I started with making a pie chart for class rank. The data given from collegeboard.com said that 14% of students were in the top 10th of the graduating class, 38% were in the top quarter, and 72% were in the top half. I plugged in the numbers to make a graph, not even realizing at first that the numbers didn't add up to 100%. After working with the numbers, I realized that I had to do some math to make everything add up. There was 14% in the top tenth, which I subtracted from the 38% in the top quarter (because if you're in the top tenth, then you are also in the top quarter) to get 24%. I then subtracted the 24% from the 72% in the top half of the class (because, again, if you're in the top quarter, you are also in the top half). This left me with 90%. I concluded that the remaining 10% were students in the bottom half of the class.

The third data set was much simpler. 91% of first year students were out of state, while 9% were in-state.

For my first graph, I tried overlapping information. In my head it seemed like a great idea, but it ended up being a confusing visual disaster, in my opinion. I started with the original colorful pie chart. I then took the class rank pie chart and filled each slice with a different pattern. The pattern was transparent, so you could still see the colored chart underneath. The last chart was much simpler, with only two pieces of data, so I created a border around the entire graph. It was really busy and the patterns made it difficult to understand the chart underneath. I had to use keys for everything, because overlapping made it hard to label everything directly.





It took me forever to figure out another solution. My AHA! moment was to make everything 3-D. I made the 3 pie charts 3 dimensional and had them somewhat stacked. However, I wanted to label everything, so the graphs are more spread out than I originally thought. This display is much easier to understand because everything is color coded and the viewer can understand the information much quicker.




In the end I was happier with the second display, which has all the elements I was hoping to include. It has three variables that tell a story about first year RWU students in a visual way. It may not be the best example of a multivariate display, but I learned a lot through the problem solving I had to do to get to my end result.