How to make Choropleth with Plotly

Resources

First, let us start with some helpful manual webpages:

1. plotly Choropleth in general.
2. Colorbar parameters.
3. Layout.
4. Trace.
5. Color scale.

Example

Here is a minimal(ish) working example:

I cut it up in parts to provide some explanation in between, I will link you to the full example code and data files by the end of this tutorial.

import pandas as pd
import plotly.express as px
import json
from urllib.request import urlopen

Plotly Choropleth works with pandas data frame, so if you have data in another form, you need to organize it in pandas.dataframe format (tutorial here).

Now, let us load the data: IMPORTANT: When loading the data frame, the fips column must be read as string (dtype) or otherwise the leading zeros will disappear.

A minimal working data frame for choropleth only need two columns:

1. a fips column, specifying locations;
2. a val column, specifying the colors of the heat map;

df = pd.read_csv('choropleth_for_wiki.csv', dtype={'fips': str})

Now we get the geojson file that contains the polygons of county boundaries.

counties_webpage = 'https://raw.githubusercontent.com/plotly/datasets/master/geojson-counties-fips.json'
with urlopen(counties_webpage) as response:
    counties = json.load(response)

Now we plot. In this example, we plot all the US counties (including Alaska and Hawaii).

fig = px.choropleth(
    df, 
    geojson=counties, 
    locations='fips', 
    color='val',
    color_continuous_scale='viridis',               # see color scale link on the top for other colormap
    range_color=(df['val'].min(), df['val'].max()), # the min max values of the color range
    scope='usa',                                    # must have or it will show a world map instead of just US counties
)

I got terribly confused at the beginning, so here is a little reminder: The fig object above has nothing to do with a Matplotlib figure object although it is also conventionally named as fig. Fundamentally, Plotly figure object is designed for interactivity (zooming in and out, moving, and showing information when pointing your mouse to a part of the map).

Now we update the layout of the figure object fig. There are a lot more parameters then shown here that you can modify (See the layout and colorbar parameters links on the top).

fig.update_layout(
    margin={'l': 0, 'b': 0, 'r': 0, 't': 0},            # can only have >=0 values
    coloraxis_showscale=True,                           # weather to show colorbar or not
    coloraxis_colorbar={
        'title': 'Risk',                                # colorbar title
        'titlefont': {'size': 25, 'color': 'black'},    # title font size and color
        ##=================== Shape ===================##
        'lenmode': 'pixels',                            # the unit of color bar height (['fraction', 'pixels'])
        'len': 500,                                     # color bar height (in this case, 500 pixels)
        'thicknessmode': 'pixels',                      # the unit of color bar width (['fraction', 'pixels'])
        'thickness': 15,                                # color bar width (in this case, 20 pixels)
        ##================ Positioning ================##
        'xpad': 25,                                     # sets the amount of padding (in px) along the x direction
        'x': .91,                                       # the x relative position ([-2,3]) of the colorbar
        ##============== Tick parameters ==============##
        'tickfont': {'size': 25, 'color': 'black'},     # tick font size and color
        'ticks': 'inside',                              # the little ticks can stick out or inside
        'ticklen': 14.5,                                # the length of the ticks
        'thicknessmode': 'pixels',                      # the unit of color bar width
    }
)

We can also change the county boundary parameters by update traces.

fig.update_traces(marker_line_width=.1)

Now we save the figure if needed.

fig.write_image(
    file='choropleth_for_wiki.pdf',             # always save as pdf for latex use
    format='pdf',                               # there are other formats, such as `png` and `jpeg`, etc.
    engine='kaleido',                           # There are other engines, but this one works for midway
    width=1000,                                 # in pixels. The bigger, the better resolution
    height=600,                                 # in pixels. The bigger, the better resolution
)

I found out that we must remove engine='kaleido' line for it to run on my laptop. But YMMV.

Since we sometimes just want to use the plot for publication, and hence we may want to disable the interactions since interaction can be a little.

disable_interaction = True
if disable_interaction:
    config = {'scrollZoom': False, 'displayModeBar': False}
else:
    config = None
fig.show(config=config)

Downloads

• Here is the code: Choropleth_example.py
• Here is the data: Choropleth_example.csv