From CS 294-10 Visualization Sp10

Jump to: navigation, search

With this visualization, I have decided to sacrifice the total amount of encoded data for the sake of trying to tell a clearer story. I wanted to focus on the fact that Haiti hasn't received nearly enough aid money compared to what it actually needs. This data set shows us that there are ten sectors that all need aid money, and none of these ten have received half of the aid needed. Some haven't received any aid at all. I have omitted the sources of the money, as well as the channel it came through.

During my design, I wanted to maximize the ability to compare the amount of money received to the amount of money needed--so I tried to choose the most effective encoding for quantitative data. I tried to use type-1 position (Cleveland and McGill), by using adjacent bars for direct comparison. Color I tried to keep consistent with generally understood schemes, where green was money received and red was money needed. I tried to keep the scale of number on a manageable level, by representing numbers in millions of dollars. I tried to keep my data-ink ratio high by using dashed lines instead of solid ones.

When it came to sorting the ten sectors, I wasn't sure if I wanted to sort it by received aid, needed aid, or by % of needed aid received (confusing, right?). I looked at all three, and sorting by needed aid (smallest to largest) was the most inoffensive. The last design decision I made was to include the raw values of each bar on the image, because three of the sectors have nearly zero received aid, I didn't want it to be confusing why some bars appeared missing. The program I was using didn't let me selectively show these values, otherwise I would have left off all non-zero values. Perhaps the class can let me know if they think it helps clarify the visualization.

I used Office 2008 for Mac (gasp!) to make this. Originally I tried to use Tableau, but I couldn't get it to work quite the way I wanted it to work, though it did make for easier initial exploration of the data.

[add comment]
Personal tools