Closed trvrb closed 7 years ago
The plan now is to table this for the time being. I'm leaving the issue open now, but will adjust text to remove references to this.
As of a297a920e011341f50a064303d7b85e4ee7bdba8, calling this resolved for the time being.
@evogytis and @maxbiostat ---
I think there is an obvious reviewer question about whether human cluster sizes show seasonality. You nicely show seasonality in cluster arrival times suggesting changes in force of infection from camels to humans through time. The corollary would be whether R changes in humans through time. Can you do a similar analysis looking at relative size of clusters in each month? I'm imagining this would look similar to Figure 2A, but instead showing "odds ratio for cluster size" on the y.
Separately, I think that this paragraph would be more easily explained if you flipped the ordering of Figure 2. Ie:
Figure 2A: Scatterplot of date vs cluster size (currently Figure 2B). Figure 2B: Violin plot of month vs odds ratio for introduction (currently Figure 2A). Figure 2C: Violin plot of month vs odds ratio for cluster size.
I always like showing the rawer data first to give people a better idea of how you're boiling things down to summaries. The scatterplot gives an easy introduction to the paragraph of "We extracted introduction times and sizes from the posterior...."