Changing legend in plotnine

Question

I have a question regarding legend in a plotnine plot.

import pandas as pd
import numpy as np
from pandas.api.types import CategoricalDtype
from plotnine import *
from plotnine.data import mpg
%matplotlib inline


c= pd.read_excel("cenpv.xlsx")
c.head()


dodge_text = position_dodge(width=0.9)

(ggplot(c, aes(x='exon', y='mean'))
 + geom_bar(stat='identity', position='dodge', show_legend=False)
 + geom_text(aes(label='percentage'),                                   
             position=dodge_text,
             size=8, va='bottom', format_string='{}%')
 + geom_hline(aes(yintercept = "Overall mean", color="Overall mean")))

I expected that legend will have just a yellow line with label overall mean. Is it possible to change it?

score 2 · Accepted Answer · answered Jan 15 '20 at 09:51

We make something that looks like your data:

c = pd.DataFrame({'exon':['CENPV_'+str(i+1) for i in range(5)],
                 'mean':np.random.poisson(100,5),
                 'percentage':np.random.randint(low=10,high=100,size=5)})
c['Overall mean'] = c['mean'].mean()

You have overall mean as a column, so ggplot2 (or plotnine) interprets it as a range of continuous values to plot the colors over.

What you need to do is to provide the mean as an array and the color as a list:

dodge_text = position_dodge(width=0.9)

(ggplot(c, aes(x='exon', y='mean'))
 + geom_bar(stat='identity', position='dodge', show_legend=False)
 + geom_text(aes(label='percentage'),                                   
             position=dodge_text,
             size=8, va='bottom', format_string='{}%')
 + geom_hline(aes(yintercept = c['mean'].mean(), color=["Overall mean"]))
 + scale_color_manual(values="yellow",name=' ')
)

Changing legend in plotnine

1 Answers1