Add demo notebook with with geographical data for various locations across the US

[andersy005]

Cc @orianac

[andersy005]

@orianac,

to rank the locations, here is the code snippet i used:

risk_df['present_rank'] = risk_df['present_risk'].rank(ascending=False).astype(int)
risk_df['future_rank'] = risk_df['future_risk'].rank(ascending=False).astype(int)

does this look match what you were envisioning?

[orianac]

these are so great @andersy005 ! putting a few requests here!
for 100-locations notebook:

• can we add two more maps to the map figure such that it is a three-panel map figure, with present day, future, and change (future - present day). this can replace the scatter plot which is a little harder to read.
• add figures representing fire risk in % (and update colorbar to "Annual fire risk [%]" and title to "Fire risk to 100 locations across United States". put the plots in % first, followed by the plots about scores.
• add subplot titles of "Current", "Future (~2047, RCP8.5)", "Difference (future - current)"
• repeat the three-panel map figure above but instead of displaying the locations with risk (%), plot the locations as scores. (e.g. translate the % into the score using the mapping outlined in the score_bins.ipynb notebook). make appropriate colorbar labels/titles.
• repeat the three-panel map figure above but instead plot the rank among the different locations. this can be a complement to the bar chart, allowing a user to see on the map which locations rank highest with risk. appropriate titles/colormaps. put these new maps above the bar chart, so that the bar chart is last.
• update the rank calculation to be based upon % risk (not score) to remove ties
• for all of the maps make the colorbar not overlap any of the maps
• cut the bar chart
• time permitting: add a bit of context (titles above different sections)

[orianac]

for transmission line map:

• make map according to score (not % risk) for both score 9 (1%) and 10 (3%). three panel map: current, future, difference
• repeat the three panel map figures above for both thresholds (9 and 10) but instead of plotting the count of pixels plot the % of pixels across the length of the line (what portion of the transmission line is in high risk areas)
• debug such that the same lines are plotted on the count and percentage plots. the only thing that should be different is the colormap/label (because in one we're plotting the count of risky pixels and in the other we're normalizng that count by the total count)