Set color for missing values in folium choropleth

Question

I have a dataframe with some countries and variables and I would like to produce a choropleth map with folium, using a geojson file for the entire world. I have a problem with folium assigning maximum value on a color scale to countries that are not present in my dataframe. Minimum exaple below:

import random
import pandas as pd
import folium
import json

map_data = pd.DataFrame({
    'A3':['POL', 'CZE', 'SVK', 'HUN', 'AUT'],
    'value':random.sample(range(10), 5)
})

m = folium.Map(
    location = [50, 15], 
    zoom_start = 4
)

m.choropleth(
    geo_data = 'https://github.com/simonepri/geo-maps/releases/download/v0.6.0/countries-land-10km.geo.json',
    data = map_data,
    columns = ['A3', 'value'],
    key_on = 'feature.properties.A3',
    fill_color = 'YlOrRd'
)

My question is the following: How can I tell folium to assign a specific color (e.g., gray or transparent) to missing countries (i.e., those present in json file but not in map_data), instead of coloring them as a maximum value for given variable (which is a strange behavior)?

Answer 1

It seems there in no way to achieve that with choropleth method. I found a workaround with custom style_function and GeoJson instead of using choropleth:

import random
import pandas as pd
import folium
from branca.colormap import LinearColormap
import json

map_data = pd.DataFrame({
    'A3':['POL', 'CZE', 'SVK', 'HUN', 'AUT'],
    'value':random.sample(range(10), 5)
})

map_dict = map_data.set_index('A3')['value'].to_dict()

color_scale = LinearColormap(['yellow','red'], vmin = min(map_dict.values()), vmax = max(map_dict.values()))

def get_color(feature):
    value = map_dict.get(feature['properties']['A3'])
    if value is None:
        return '#8c8c8c' # MISSING -> gray
    else:
        return color_scale(value)

m = folium.Map(
    location = [50, 15], 
    zoom_start = 4
)

folium.GeoJson(
    data = 'https://github.com/simonepri/geo-maps/releases/download/v0.6.0/countries-land-10km.geo.json',
    style_function = lambda feature: {
        'fillColor': get_color(feature),
        'fillOpacity': 0.7,
        'color' : 'black',
        'weight' : 1,
    }    
).add_to(m)

Answer 2

This has been fixed in folium in a recent pull request: https://github.com/python-visualization/folium/pull/1005

If you install folium from git (or from PyPY once the 0.7 version is released) you can use the nan_fill_color and nan_fill_opacity arguments of the choropleth method of Map to style elements without a value.

The final example in this Notebook shows how to do that: https://nbviewer.jupyter.org/github/python-visualization/folium/blob/master/examples/GeoJSON_and_choropleth.ipynb#Using-choropleth-method