Mashups 7: Corruption
This week’s assignment was to finish project 2, which combines at least one API and at least one library. So I present to you… UAE historical politics!
I wanted to learn how to use D3 to do data visualization, and relatively easy dataset to use was UAE Politics from Quandl. I personally thought that it was an interesting, serious subject, and also one not really discussed (that should be discussed more) in this country.
I chose to use the delimited multi-series chart for this dataset, since it was clean, versatile, and had a wide variety of types of graphs to explore the data. The colors were rather untasteful, but that could be easily changed. In addition the documentation was solid and everything was uploaded to github. I was particularly impressed by the interactivity of the hover over functionality of each of the data points.
The only trouble I had creating the first prototype was just syncing up the data to the format of the original. I had my APIs from five separate csv files and needed to combine them to one. I originally used a streamograph for the data, but after some feedback many people found it too confusing:
In addition the negative data points were bugging the program. I decided that even though a line graph has less wow factor, it serves as a much better representation for its purposes. Another point of feedbcak is that people didn’t understand what the data points meant or where it was coming from. Hence I used a free library to create a tabbed menu for each of the variables at the bottom. Then I got feedback that it would be better to color code the menu and delete the legend, since it was an eyesore.
The last part was cleaning up the presentation by changing to custom Google fonts, choosing a bright, beautiful color scheme, and making sure that everything was centered and legible.
After finishing the first prototype, I received some great feedback from the class. The aesthetics were great, but the website lacked purpose and direction. To me, the message of the data was obvious: the UAE is a very politically stable country, scoring high on all variables except for voice and accountability, in which the UAE government has a tight rein on free press. However, to my audience, the message was not obvious. The website needed more context for someone who has little knowledge of UAE political data to really engage with, or more meat for someone with a lot of knowledge of UAE politics to explore with.
This Quadrigram article is a great tutorial for teaching the ropes for data visualization, and I wish I had discovered it earlier. Let’s break them down one by one.
This data visualization was originally created for exploratory, not explanatory purposes. This was mainly because since I am not an expert on the subject matter, this was an additional resource for those knowledgeable in the field to see clean data in an uncluttered, unbiased manner. I intentionally did not want any narrative bias when an academic explores the data to reach his or her own conclusions.
Unfortunately, for this class, the explanatory approach would in retrospect make a more engaging project since the audience that will actually see it is only the other students in the class. In that sense I should have gone through Quadrigram’s step by step process. An alternative approach to this project could be:
- Define the goals — Educate the everyday resident of the UAE its political history.
- Know your readers — Everyday resident of the UAE, or well-educated observer with interest in UAE politics — most likely white-collared expats, perhaps some local Emiratis as well
- Data, what type is it? How many dimensions? What variables? — A combination of quant and qualitative data would be best. I would need additional sources for qualitative data, and I would need to find the appropriate D3 library to combine both.
