leahannejohnson
commented
3 years ago Update 2:
By this week, we were hoping to have a wrangled data set, which we do. Thai and I were fairly busy this week, so Sam took the lead on that and also got us a bit ahead of schedule by doing some preliminary spatial visualizations!
In terms of next week, we're on track. Thai and I will work more on the visualizations, and I think there's a way we can get our data into GitHub that I'll look into more over the weekend as well. Apart from that, we still need to add the network visualizations and polish everything up for the blog post. We hope to have a rough version by midweek, which we can then refine further in the remaining time.
katcorr
samassaf
Sam Leah Thai Not Stonks
In this project, we will shift our focus to the exploration of international trade relationships. Trade between countries plays a huge role in the development of several sectors of their individual economies, creating millions of jobs worldwide. In particular, we would like to visualize these relationships, and identify natural groupings among countries, as well as figure out how well-connected different countries are.
In order to do so, we will download data from CEPII - specifically the trade flows data set, which incorporates yearly bilateral trades down to the product level. Products are grouped using the Harmonized System (HS), this is the standard nomenclature for international trade, more simply stated, this is the base for classifying different products that are traded. The HS has been updated 6 times, all of which are accessible on this website. The updated HS can only be applied to the trade flow data after it is created. The data set includes variables such as year, product category (HS code), exporter, imported, value of the trade (in $1,000), and lastly quantity (in metric tons).
We hope to have at least two visualizations, with an interactive component if possible. The first is a visualization of the international trade network, weighted in some manner to enable us to identify well-connected countries (though we have yet to identify the measure of connectivity we will consider here). The second we envision to be some sort of choropleth with similar information, either identifying groups of related countries, or how well countries are connected. Looking to identify natural groupings among the countries, we would also like to incorporate clustering, and as such might include a dendrogram if a hierarchical approach is used. We have yet to decide how interactivity will be included, but this might take the form of a hover or click feature similar to that in our Shiny app from the previous project.
In terms of a timeline, we will meet at least bi-weekly on zoom to catch up in “person”, as well as communicating over GroupMe. We plan to have a wrangled data set by mid next week, and a working blog post (with rough drafts of each visualization) a week from then. From then on, we will focus on cleaning up the post, and including any interesting features we might see fit given our analysis.