How to annotate grouped bars with the value and text from a different column

Question

I am trying to get a bar plot in seaborn with vertical text alignment. The name text is aligned vertically but it maintains some gap between each bars. I need to align the name text in top of the bar vertically and also the col3 values inside the bar. Please help me out to solve this.

The code that I had tried is,

df= pd.DataFrame({'name':['Name1','Name2','Nmae3','Name4','Name2','Name1'],'Col1':[1,1,1,2,2,2],'col2':['a','b','c','a','b','c'],'col3':[22,33,4,11,3,43]})

    name  Col1 col2  col3
0  Name1     1    a    22
1  Name2     1    b    33
2  Nmae3     1    c     4
3  Name4     2    a    11
4  Name2     2    b     3
5  Name1     2    c    43

tm = sns.barplot(x="Col1", y="col3", hue="col2", data=df)

for index in range(len(df)):
    tm.text(index,df.col3.values[index]+0.2, str(df.name.values[index]),fontdict=dict(color='black', rotation=90,fontsize=10),horizontalalignment='left')

Answer 1

• The main issue is getting the correct 'name' value from the dataframe.
• Use matplotlib.pyplot.bar_label and custom labels, as described in:
  • How to plot and annotate grouped bars in seaborn / matplotlib
  • How to add value labels on a bar chart
• Use the column with the hue value ('col2') and the height of the bar to get the correct 'name' annotation for the top of the bar.
  • df.loc[(df.col2.eq(col) & df.col3.eq(h)), 'name'].iloc[0]
  • Both columns are used to get the correct 'name' incase the same value is in multiple groups.
• Grouped bars are plotted by the order of the hue groups, so both 'a' bars, then 'b' bars, etc.
• Tested in python 3.10, pandas 1.4.3, matplotlib 3.5.1, seaborn 0.11.2
  • Assignment expression (:=) - python >= 3.8
  • .bar_label - matplotlib >= 3.4.0

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(9, 7))
sns.barplot(x='Col1', y='col3', hue='col2', data=df, ax=ax)

# get the unique values for hue
hue_col = df.col2.unique()

# iterate through each group of containers - in order by hue groups
for c, col in zip(ax.containers, hue_col):
    
    # use the column and bar height to get the correct value for name
    labels = [f"{df.loc[(df.col2.eq(col) & df.col3.eq(h)), 'name'].iloc[0]}" if (h := v.get_height()) > 0 else '' for v in c ]
    # labels without using assignment expression
    # labels = [f"{df.loc[(df.col2.eq(col) & df.col3.eq(v.get_height())), 'name'].iloc[0]}" if v.get_height() > 0 else '' for v in c ] 
 
    # add the name annotation to the top of the bar
    ax.bar_label(c, labels=labels, padding=3)  # rotation=90 if needed
    
    # add the bar value annotation inside the bar
    ax.bar_label(c, label_type='center')

    # pad the spacing between the number and the edge of the figure
    ax.margins(y=0.1)