datalorax
commented
5 years ago Nice review Jon,
Ouafaa - if you decide to go forward with the geographic map I'd really encourage you to read the chapter on maps from Wilke's book. It's pretty great.
I agree that ggarange is really nice. I usually use {cowplot} or {patchwork}, but this looks nice too.
Ouafaa, wow, what a great portfolio! I especially appreciated the way that you structured your visualization order, where a clear story was told by beginning with that global view of Kiva frequency loans and then becoming more specific with each of the subsequent graphs. I will also admit that I am currently obsessed with any sort of geographic map, so as soon as I saw the first global map, I was completely sold! Although you do not have the map listed as one of your three official visualizations, I think it has a lot of potential to complement the bar graph that lists number of loans borrowed by world region, however I think it would be useful to break up the map into each region and present the results in some sort of wrapped fashion, because right now it is difficult to get a complete global picture. Your Visualization #1 is such a great grouping of graphs. I had no idea that you could combine multiple graphs into one visualization and I am definitely going to use ggarrange() for my own portfolio as I think it provides a type of static data dashboard where the viewer can immerse themselves in the multiple visualizations simultaneously. One change that may help clarify each graph's message is re-ordering the graphs so that the regions and top 10 countries bar graphs are both presented on the top together, and then you move into the next two graphs on the bottom row with the sectors being displayed together. Really solid Visualization #2, I agree that the word cloud provides contextual information that a bar graph cannot convey. Titles for each cloud would really help better understand the different conditions that you are comparing. On your Visualization #3, have you attempted using fill = world_region instead of facet_wrap and assigning an alpha? It might be too much for one graph, but if you are committed to the geom_point than it might be another option to try.
Thanks again for sharing! @datalorax