How to plot a bar plot of 2 categorical columns using matplotlib or seaborn

Question

This might be a simple task but I am new to plotting in python and is struggling to convert logic into code. I have 2 columns like below. 0 mean not churned and 1 means churned. gender is an object column and churned is a category column

gender|churned
--------------
male   |0
male   |1
female |0
female |1
female |1
male   |1

I simply want a stacked bar graph (please correct me if this is not the right choice of graph) with 0 and 1 on x axis (churn column) and for each of those 2 categories I want a stacked bar graph with 2 different colours for each gender showing the total number of males and females under 0 (not churned) and total number of males and females under 1(churned).

I tried:

df.Churn.value_counts().plot(kind='bar') 

it gave me the total count for each 0 and 1 but i need it divided by gender aswell.

Hope I am making sense

Answer 1

You can table it:

import pandas as pd
df = pd.DataFrame({'gender':['male','male','female','female','female','male'],
                'churned':[0,1,0,1,1,1]})
pd.crosstab(df['churned'],df['gender']).plot(kind="bar",stacked=True)

Answer 2

If you wanted an interactive version then you could use hvplot:

import pandas as pd
import hvplot.pandas #noqa

# 1. CREATE DF:
churn = pd.DataFrame({"gender":["male","male","female","female","female","male"],
                     "churned":[0,1,0,1,1,1]})
churn

Out[2]: 
   gender  churned
0    male        0
1    male        1
2  female        0
3  female        1
4  female        1
5    male        1

# 2. GROUP THE DATA BY "churned" THEN "gender":
plot_me = churn.groupby(["churned","gender"])[["gender"]].count().rename(columns={"gender":"count"})
plot_me

Out[3]: 
                count
churned gender       
0       female      1
        male        1
1       female      2
        male        2

# 3. PLOT:
plot_me.hvplot.bar(stacked=True,color=["maroon","teal"],line_width=3,
                   line_color="black",height=350,width=500)

Out[4]: