Adding outline stroke to plotnine annotations¶

When you plot text on top of busy charts, sometimes the text blends in with the details in the chart underneath. In this guide we'll use matplotlib's text_effects to make text on top of a chart readable.

Read in our data¶

We have two datasets: the original data, and the annotations we want to put on the graphic.

In [6]:

import pandas as pd
from plotnine import *

df = pd.read_csv("countries.csv")
df.head()

import pandas as pd from plotnine import * df = pd.read_csv("countries.csv") df.head()

Out[6]:

	country	continent	gdp_per_capita	life_expectancy	population
0	Afghanistan	Asia	663	54.863	22856302
1	Albania	Europe	4195	74.200	3071856
2	Algeria	Africa	5098	68.963	30533827
3	Angola	Africa	2446	45.234	13926373
4	Antigua and Barbuda	N. America	12738	73.544	77656

In [7]:

annotations = df[df.country.isin(['Tonga', 'Angola', 'Iceland'])]
annotations

annotations = df[df.country.isin(['Tonga', 'Angola', 'Iceland'])] annotations

Out[7]:

	country	continent	gdp_per_capita	life_expectancy	population
3	Angola	Africa	2446	45.234	13926373
74	Iceland	Europe	31092	79.720	281210
169	Tonga	Oceania	4886	70.810	97935

The problem¶

When you plot text on top of busy charts, sometimes the text blends in with the details in the chart underneath.

In [8]:

(
    ggplot(df)
    + aes(x='gdp_per_capita', y='life_expectancy', label='country')
    + geom_point(color='black')
    + geom_point(annotations, color='red')
    + geom_text(annotations, ha='left', nudge_x=1000)
)

( ggplot(df) + aes(x='gdp_per_capita', y='life_expectancy', label='country') + geom_point(color='black') + geom_point(annotations, color='red') + geom_text(annotations, ha='left', nudge_x=1000) )

Out[8]:

<ggplot: (312649858)>

To help the text stand out, we will add an outline to it.

Adding stroke to text annotations¶

To add an outline to our text, we will use the matplotlib's path effects. The simplest way is using withStroke.

In [26]:

import matplotlib.patheffects as path_effects

(
    ggplot(df)
    + aes(x='gdp_per_capita', y='life_expectancy', label='country')
    + geom_point(color='black')
    + geom_point(annotations, color='red')
    + geom_text(annotations,
                ha='left',
                nudge_x=1000,
                path_effects=[
                    path_effects.withStroke(linewidth=3, foreground='white')
                ])
)

import matplotlib.patheffects as path_effects ( ggplot(df) + aes(x='gdp_per_capita', y='life_expectancy', label='country') + geom_point(color='black') + geom_point(annotations, color='red') + geom_text(annotations, ha='left', nudge_x=1000, path_effects=[ path_effects.withStroke(linewidth=3, foreground='white') ]) )

Out[26]:

<ggplot: (311895471)>

By giving it a linewidth and a foreground, you get a certain thickness of border around your text label. Now you can read it!

In [ ]: