Ahmed Elbashir & Belinda Shi
Our group was on the small side and we did our best to split the work evenly among the 2 of us. Ahmed wrote a script for downloading and cleaning the data while Bee prepared the scrollytelling code. Each person took primary responsibility in creating 2 of the 4 visualizations, and later improvements to visualizations, such as adding axis labels or resizing was evenly split. Work sessions were synchronized and typically took the form of working together for an entire weekend, consulting with each other regularly on difficult decisions and to avoid troublesome merge conflicts. We also split writing our deliverables roughly evenly, and each of us took responsibility for different sections of our final paper and of our other deliverables.
The research and development process met with more hiccups at first as we attempted to make sense of data discrepancies and figure out basic bugs in our scrollytelling code and map projection code. After acquiring the data, we explored for a while in Tableau (discovering and fixing the discrepancies along the way), making prototypes of our visualizations when we discover an interesting phenomena and a good way to visualize it, and exporting the data Tableau uses for the visualization so that we can visualize it in d3 without writing complicated transformations in Javascript. Writing the visualizations in d3 required a lot of bug fixing, googling errors and iteration on our designs, and the final product was roughly a thousand lines of code.