Prevent overlapping labels in plotnine with automatic positioning¶

To prevent labels from overlapping on top of each other, ggplot has ggrepel, while plotnine has adjustText.

Read in our data¶

In [2]:

import pandas as pd
from plotnine import *

df = pd.read_csv("countries.csv", nrows=10)
df.head(2)
import pandas as pd
from plotnine import *

df = pd.read_csv("countries.csv", nrows=10)
df.head(2)

Out[2]:

	country	continent	gdp_per_capita	life_expectancy	population
0	Afghanistan	Asia	663	54.863	22856302
1	Albania	Europe	4195	74.200	3071856

The problem¶

We have labels that conflict with both points and other labels. They are drawn on top of each other, which isn't ideal.

In [3]:

(
    ggplot(df)
    + aes(x='gdp_per_capita', y='life_expectancy', label='country')
    + geom_point()
    + geom_text(size=10)
    + ylim(40, 90)
    + xlim(0, 35000)
)
(
    ggplot(df)
    + aes(x='gdp_per_capita', y='life_expectancy', label='country')
    + geom_point()
    + geom_text(size=10)
    + ylim(40, 90)
    + xlim(0, 35000)
)

Out[3]:

<ggplot: (312224603)>

The solution¶

You'll need to pip install adjustText for this to work (and maybe restart your kernel for it to pick up the install). By passing a list of adjust_text options to geom_text, plotnine will reposition labels until it is "right" (and usually not overlapping).

In [21]:

# Text elements avoid each other
# Text elements avoid being drawn on top of points 
# Bounding box for text is expanded 1.1x width 1.25x height

adjust_text_options = {
    'avoid_text': True,
    'avoid_points': True,
    'expand_text': (1.1, 1.25),
}

(
    ggplot(df)
    + aes(x='gdp_per_capita', y='life_expectancy', label='country')
    + geom_point()
    + geom_text(adjust_text=adjust_text_options, size=10)
    + ylim(40, 90)
    + xlim(0, 35000)
)
# Text elements avoid each other
# Text elements avoid being drawn on top of points 
# Bounding box for text is expanded 1.1x width 1.25x height

adjust_text_options = {
    'avoid_text': True,
    'avoid_points': True,
    'expand_text': (1.1, 1.25),
}

(
    ggplot(df)
    + aes(x='gdp_per_capita', y='life_expectancy', label='country')
    + geom_point()
    + geom_text(adjust_text=adjust_text_options, size=10)
    + ylim(40, 90)
    + xlim(0, 35000)
)

Out[21]:

<ggplot: (315228046)>

I honestly don't think it works very well but it's The Suggested Solution.

In [ ]: