Figure out how to visualize uncertainty

[cypressf]

I made a mockup of how this could work, for both national, and county-level projections. This requires a predefined "high risk" and "low risk" threshold for each variable so we can color the map differently year-over-year, instead of marking the 100th percentile as dark red and the 0th percentile dark blue.

[cypressf]

On a more specific projection visualization, we've done some work with Climate Moisture Index and asked questions about how we can best visualize the change over time, and the likelihood of different changes occurring. We've questioned what is useful to highlight on the map, for example do we care if a dry spot gets even dryer, if a wet spot gets dryer, if a wet spot gets wetter?

[cypressf]

Jen asks: How many scenarios are we thinking of running? Should we have 4 universal scenarios for all metrics, or custom scenarios for each metric?

John says: the focus of the website could just be the change from now to 2040-2060 avg, instead of showing all the dates. In the future, will it be worse or better? A map showing how much more wet or dry instead of showing the graph of the change over time.

With CMI, we did that, but in order to color the risks, we tried to show changes that jumped outside of a category: substantially wetter than it was, substantially dryer than it was.

[cypressf]

Let's make this projection site a different entry-point from the main triage platform for now because the data is going to look different between model-generated projections and the historical data we have now.

[cypressf]

For future projections:

• divide each data set into 5 categories, from very low to very high risk
• within each category, we linearly map the data from 0 - 20, 20 - 40... 80 - 100