Text annotations for means and medians on a plotnine graphic¶

We'll start by reading in our dataset. It's a lot of data, so we probably want to add some nice summaries for people to understand what's going on.

In [37]:

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[37]:

	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

Using after_stat for text labels¶

To add a point or line to a chart that represents mean or median, you simply use stat_summary. It takes x and y right out of aes.

If you want to label the point, though, it's more difficult. Normally aes points to a column, but we want a calculated point, the median. It doesn't exist in the dataset! To solve this problem, we use after_stat.

In [22]:

import numpy as np

(
    ggplot(df)
    + aes(x='continent', y='life_expectancy')
    + geom_point(alpha=0.5, color='grey')
    + stat_summary(geom='point', fun_y=np.median, color='red')
    + stat_summary(aes(label=after_stat('y')), geom='text',  fun_y=np.median)
)

import numpy as np ( ggplot(df) + aes(x='continent', y='life_expectancy') + geom_point(alpha=0.5, color='grey') + stat_summary(geom='point', fun_y=np.median, color='red') + stat_summary(aes(label=after_stat('y')), geom='text', fun_y=np.median) )

Out[22]:

<ggplot: (314219617)>

Let's break down a few parts of that last line:

• stat_summary(: instead of a column, we're plotting a statistical summary
• label=after_stat('y'): set the label (the displayed text) to be the y position after we calculate the statistics
• geom='text': instead of a point or line, we want text for each statistical summary
• fun_y=np.median: calculate the median!

Rounding after_stat median annotations¶

The numbers that come out of after_stat look awful, with many numbers after the decimal point. Instead of using the raw numbers, use after_stat('round(y)') to round them! And yes, round goes inside of the string.

In [36]:

import numpy as np

(
    ggplot(df)
    + aes(x='continent', y='life_expectancy')
    + geom_point(alpha=0.5, color='grey')
    + stat_summary(geom='point', fun_y=np.median, color='red')
    + stat_summary(
        aes(label=after_stat('round(y, 2)')),
        geom='text',
        fun_y=np.median,
        nudge_x=0.075,
        size=10,
        ha='left')
)

import numpy as np ( ggplot(df) + aes(x='continent', y='life_expectancy') + geom_point(alpha=0.5, color='grey') + stat_summary(geom='point', fun_y=np.median, color='red') + stat_summary( aes(label=after_stat('round(y, 2)')), geom='text', fun_y=np.median, nudge_x=0.075, size=10, ha='left') )

Out[36]:

<ggplot: (312551310)>

In [ ]: