In [1]:
import os
import folium

print(folium.__version__)
0.5.0+105.g065f6f3.dirty

Using folium.colormap

A few examples of how to use folium.colormap in choropleths.

Let's load a GeoJSON file, and try to choropleth it.

In [2]:
import json
import pandas as pd

us_states = os.path.join('data', 'us-states.json')
US_Unemployment_Oct2012 = os.path.join(
    'data', 'US_Unemployment_Oct2012.csv'
)

geo_json_data = json.load(open(us_states))
unemployment = pd.read_csv(US_Unemployment_Oct2012)

unemployment_dict = unemployment.set_index('State')['Unemployment']

Self-defined

You can build a choropleth in using a self-defined function. It has to output an hexadecimal color string of the form #RRGGBB or #RRGGBBAA.

In [3]:
def my_color_function(feature):
    """Maps low values to green and hugh values to red."""
    if unemployment_dict[feature['id']] > 6.5:
        return '#ff0000'
    else:
        return '#008000'
In [4]:
m = folium.Map([43, -100], tiles='cartodbpositron', zoom_start=4)

folium.GeoJson(
    geo_json_data,
    style_function=lambda feature: {
        'fillColor': my_color_function(feature),
        'color': 'black',
        'weight': 2,
        'dashArray': '5, 5'
    }
).add_to(m)

m.save(os.path.join('results', 'Colormaps_0.html'))

m
Out[4]:

StepColormap

But to help you define you colormap, we've embedded StepColormap in folium.colormap.

You can simply define the colors you want, and the index (thresholds) that correspond.

In [5]:
import branca.colormap as cm


step = cm.StepColormap(
    ['green', 'yellow', 'red'],
    vmin=3, vmax=10,
    index=[3, 4, 8, 10],
    caption='step'
)

step
Out[5]:
310
In [6]:
m = folium.Map([43, -100], tiles='cartodbpositron', zoom_start=4)

folium.GeoJson(
    geo_json_data,
    style_function=lambda feature: {
        'fillColor': step(unemployment_dict[feature['id']]),
        'color': 'black',
        'weight': 2,
        'dashArray': '5, 5'
    }
).add_to(m)

m.save(os.path.join('results', 'Colormaps_1.html'))

m
Out[6]:

If you specify no index, colors will be set uniformely.

In [7]:
cm.StepColormap(['r', 'y', 'g', 'c', 'b', 'm'])
Out[7]:
0.01.0

LinearColormap

But sometimes, you would prefer to have a continuous set of colors. This can be done by LinearColormap.

In [8]:
linear = cm.LinearColormap(
    ['green', 'yellow', 'red'],
    vmin=3, vmax=10
)

linear
Out[8]:
310
In [9]:
m = folium.Map([43, -100], tiles='cartodbpositron', zoom_start=4)

folium.GeoJson(
    geo_json_data,
    style_function=lambda feature: {
        'fillColor': linear(unemployment_dict[feature['id']]),
        'color': 'black',
        'weight': 2,
        'dashArray': '5, 5'
    }
).add_to(m)

m.save(os.path.join('results', 'Colormaps_2.html'))

m
Out[9]:

Again, you can set the index if you want something irregular.

In [10]:
cm.LinearColormap(
    ['red', 'orange', 'yellow', 'green'],
    index=[0, 0.1, 0.9, 1.0]
)
Out[10]:
0.01.0

If you want to transform a linear map into a step one, you can use the method to_step.

In [11]:
linear.to_step(6)
Out[11]:
3.010.0

You can also use more sophisticated rules to create the thresholds.

In [12]:
linear.to_step(
    n=6,
    data=[30.6, 50, 51, 52, 53, 54, 55, 60, 70, 100],
    method='quantiles',
    round_method='int'
)
Out[12]:
31100

And the opposite is also possible with to_linear.

In [13]:
step.to_linear()
Out[13]:
310

Build-in

For convenience, we provide a (small) set of built-in linear colormaps, in folium.colormap.linear.

In [14]:
cm.linear.OrRd_09
Out[14]:
0.01.0

You can also use them to generate regular StepColormap.

In [15]:
cm.linear.PuBuGn_09.to_step(12)
Out[15]:
0.01.0

Of course, you may need to scale the colormaps to your bounds. This is doable with .scale.

In [16]:
cm.linear.YlGnBu_09.scale(3, 12)
Out[16]:
312
In [17]:
cm.linear.RdGy_11.to_step(10).scale(5, 100)
Out[17]:
5100

At last, if you want to check them all, simply ask for linear in the notebook.

In [18]:
cm.linear
Out[18]:
Pastel1_030.01.0
Pastel1_050.01.0
Pastel1_040.01.0
Pastel1_07