Plotly express colorscale map by absolute value instead of 0-1

September 11, 2022

I have a map of the US and am plotting winning margin in Presidential elections in Plotly Express. I want a winning margin of 0 to be displayed as white and the scale to diverge into red/blue. However, the color_continuous_scale keyword argument takes a range from 0-1, and where a winning margin of 0 falls in this 0-1 range varies each election.

Is there a workaround for this that would allow me to define 0 as white, and the extremes as red/blue?

i have tried to convert the winning margin of 0 to a fraction between 0 and 1 but this changes with every election cycle.

fig = px.choropleth(merged,
                    locationmode='USA-states',
                    locations='state_po',
                    animation_frame='year',
                    color='R_margin_x',
                    color_continuous_scale=[(0, 'blue'),
                                            (merged['R_margin_x'].min()/(merged['R_margin_x'].min()-merged['R_margin_x'].max()), 'white'),
                                            (1, 'red')],
                    scope='usa')
fig.show()

dataframe is of the form

party_simplified state_po  year   DEMOCRAT  LIBERTARIAN     OTHER  REPUBLICAN    D_pct_x   L_pct_x   O_pct_x    R_pct_x  R_margin_x    D_pct_y   L_pct_y   O_pct_y    R_pct_y  R_margin_y     R_lean
0                      AK  2008   123594.0       1589.0    7173.0    193841.0  37.889374  0.487129  2.198978  59.424520   21.535146  52.761558  0.388418  1.488455  45.361569    -7.39999  28.935135
1                      AL  2008   813479.0          NaN   19794.0   1266546.0  38.740434       NaN  0.942653  60.316913   21.576479  52.761558  0.388418  1.488455  45.361569    -7.39999  28.976468
2                      AR  2008   422310.0       4776.0   21514.0    638017.0  38.864660  0.439529  1.979906  58.715904   19.851245  52.761558  0.388418  1.488455  45.361569    -7.39999  27.251234
3                      AZ  2008  1034707.0      12555.0   16102.0   1230111.0  45.115251  0.547423  0.702079  53.635248    8.519997  52.761558  0.388418  1.488455  45.361569    -7.39999  15.919987
4                      CA  2008  8274473.0      67582.0  208064.0   5011781.0  61.012638  0.498323  1.534180  36.954859  -24.057780  52.761558  0.388418  1.488455  45.361569    -7.39999 -16.657790

>Solution :

In plotly you can pass the midpoint for continuous scales:

color_continuous_scale=px.colors.sequential.RdBu, 
color_continuous_midpoint=0.0

The RdBu scale goes from red to white to blue. By passing color_continuous_midpoint=0.0 you specify that 0.0 is in the middle of the scale, i.e. white. The other colors will be determined accordingly. The same goes if you use a custom continuous color scale.