Figure 3 comments

[nickreich]

A few general comments about figure 3:

• Some time back, @lshandross @elray1 and I had talked about having a default visualization of a pmf forecast like this being different than what is presented here. Based on focus-group results that FluSight commissioned years ago, there was a push to represent pmf forecasts with horizon on x-axis, categories on y-axis and intensity of color representing the probability. I think one could make a compelling argument to change the vis in Figure 3 to have this representation instead of the current approach (which has probability on the y-axis).
• that said, if we keep the current configuration for Figure 3, I would suggest changing the order in which the bins are stacked so that the "low" probability is at the bottom and the "very high" is at the top. Right now it is reversed, and my brain really struggles to interpret the probabilities and orderings.

[eahowerton]

@nickreich thanks for the suggestion, this seems like a much more intuitive way to visualize the pmf forecasts. Here's where I've landed:

Noting 2 choices I've made (and feel free to disagree with either one):

1. I've used a color scale where lighter colors = lower probability
2. I've chosen to keep Fig 5a as is. We could convert it to this format as well, but the main point of this panel is to show the ensemble probabilities are equivalent, and I think the new configuration would make this comparison more difficult (i.e., readers would have to compare color intensities instead of bar heights).

[nickreich]

This looks great to me!