Possible problem with geom_density() with after_stat('count')

[marcomoretto]

This code works as expected (density histogram with kde)

from plotnine import *
import pandas as pd

p_df = pd.read_csv('https://raw.githubusercontent.com/mwaskom/seaborn-data/master/penguins.csv')

(
    ggplot(p_df.dropna(), aes(x='body_mass_g', y=after_stat('density')))
    + geom_density(aes(color='species'), alpha=0.8)
    + geom_histogram(aes(fill='species'), bins=20, alpha=0.5, position='identity')
)

But switching from after_stat('density') to after_stat('count') doesn't.

from plotnine import *
import pandas as pd

p_df = pd.read_csv('https://raw.githubusercontent.com/mwaskom/seaborn-data/master/penguins.csv')

(
    ggplot(p_df.dropna(), aes(x='body_mass_g', y=after_stat('count')))
    + geom_density(aes(color='species'), alpha=0.8)
    + geom_histogram(aes(fill='species'), bins=20, alpha=0.5, position='identity')
)

The problem might be related to geom_density "ignoring" that count should be used instead of density. Same code as before with one histogram commented out.

from plotnine import *
import pandas as pd

p_df = pd.read_csv('https://raw.githubusercontent.com/mwaskom/seaborn-data/master/penguins.csv')

(
    ggplot(p_df.dropna(), aes(x='body_mass_g', y=after_stat('count')))
    + geom_density(aes(color='species'), alpha=0.8)
    # + geom_histogram(aes(fill='species'), bins=20, alpha=0.5, position='identity')
)

The documentation state that it should work as expected

[marcomoretto]

I think the problem is on my side according to this https://stackoverflow.com/questions/37404002/geom-density-to-match-geom-histogram-binwitdh

So I should rewrite the code as follow

(
    ggplot(p_df.dropna(), aes(x='body_mass_g'))
    + geom_density(aes(color='species', y=after_stat('count * 100')), alpha=0.8)
    + geom_histogram(aes(fill='species', y=after_stat('count')), binwidth=100, alpha=0.5, position='identity')
)

[has2k1]

I think the problem is on my side according to this https://stackoverflow.com/questions/37404002/geom-density-to-match-geom-histogram-binwitdh

So I should rewrite the code as follow

(
    ggplot(p_df.dropna(), aes(x='body_mass_g'))
    + geom_density(aes(color='species', y=after_stat('count * 100')), alpha=0.8)
    + geom_histogram(aes(fill='species', y=after_stat('count')), binwidth=100, alpha=0.5, position='identity')
)

Yes you have to match parameters. It makes sense but it is not obvious why. I think it merits an example in the documentation.