This is my entry for the "Storytelling With Data"-challenge, March 2019 edition. For the challenge, a dataset on aid data was chosen. The AidData research lab collected an amazing 1.5 million transactions in international aid going back to the seventies.
The challenge was to create visualizations responding to three sets of questions:
WHO DONATES? How are donations distributed across countries? Who donates to whom? Are there any patterns, for example some group of countries tends to donate only to some specific group of other countries? Or maybe some countries tend to receive only from a specific set of countries?
HOW MUCH DO THEY DONATE? How much do countries donate and receive? Who donates the most/least? Are there countries that donate and also receive? How does the amount donated/received by country change over time?
WHY DO THEY DONATE? Do countries tend to send (or receive) donations for specific reasons? For instance, is it possible that some countries tend to receive/send certain type of donations whereas other receive/send different types?
The biggest difficulty with the aid data was, that the distributions are quite skewed for both donors and recipients. Smaller recipients/donors therefore quickly tend to vanish. A second difficulty was the large number of countries.
The analysis was done in Jupyter Lab with pandas. The data was visualized in a first step with Vega-Lite and then imported into Sketch to combine everything and add the finishing touches like labels and titles.