如何 plot 带有注释的水平堆叠条

Question

• I used the example for Discrete distribution as horizontal bar chart example on matplotlib Discrete distribution as horizontal bar chart to create a chart showing share of the vote in Shropshire elections 2017.

• However, because I did not know how to manipulate the data I had to manually enter my data in the program which is clearly down to my own ignorance.

• I have the relevant data in a CSV file and can therefore load it as a dataframe.

  • The CSV has a row for each ward, of which there are 63 and columns for the % vote for each party (Tory, LD, Labour, Green, Indep) so 5 substantive columns.

• I wanted advice as to how to change the form of the data so it resembles the input for this chart.

• I am not sure what it is but seems possibly a dictionary type with key and value:

My data reads in part:

import pandas as pd
import matplotlib.pyplot as plt

category_names = ['Labour', 'LD', 'Indep', 'Green', 'Tory']
results = {'Abbey': [16, 56, 4,0, 24],
           'Albrighton': [0, 0, 32, 0, 68],
           'Alveley & Claverley': [0, 25, 0, 0, 75],
           'Bagley': [30, 30, 0, 0, 40],
           'Battlefield': [34, 0, 0, 9, 57],
           'Bayston Hill, Column & Sutton': [53, 4, 3, 7, 33],
           'Belle Vue': [43,28,0,5,24]}


# setup dataframe using the dict provided in the OP
df = pd.DataFrame(results, index=category_names)

# display(df)
        Abbey  Albrighton  Alveley & Claverley  Bagley  Battlefield  Bayston Hill, Column & Sutton  Belle Vue
Labour     16           0                    0      30           34                             53         43
LD         56           0                   25      30            0                              4         28
Indep       4          32                    0       0            0                              3          0
Green       0           0                    0       0            9                              7          5
Tory       24          68                   75      40           57                             33         24

• I am trying to get the data to be formatted like this directly from the csv file when entered as a pandas dataframe.

• Have tried the values method and the to_dict method and while they get data looking similar they are not quite correct.

  • I believe there is a need to divide the data into keys and values but that is where my knowledge hits its limits.

Answer 1

• Both options us DataFrame df from the OP.
• Use pandas.DataFrame.plot with the parameter stacked=True

Option 1: 'Party' as the y-axis

Using matplotlib from version 3.4.2

• Use matplotlib.pyplot.bar_label
  • See this answer for additional details and examples with .bar_label .
• See the matplotlib: Bar Label Demo page for additional formatting options.
• Tested in pandas 1.3.2 , python 3.8 ^1. , and matplotlib 3.4.2 ^1.
  • ^{1. Minimum version required}
  • labels = [f'{v.get_width():.0f}' if v.get_width() > 0 else '' for v in c ] without using the assignment expression ( := )
• Use .get_height() for vertical bars.

ax = df.plot.barh(stacked=True, cmap='tab10', figsize=(16, 10))

for c in ax.containers:

    # format the number of decimal places and replace 0 with an empty string
    labels = [f'{w:.0f}' if (w := v.get_width()) > 0 else '' for v in c ]
    
    ax.bar_label(c, labels=labels, label_type='center')

Using matplotlib before version 3.4.2

• Extract the .patch components in a loop, and then only plot annotations for values greater than 0.

# plot 
ax = df.plot.barh(stacked=True, cmap='tab10', figsize=(16, 10))

# annotations:
for p in ax.patches:
    left, bottom, width, height = p.get_bbox().bounds
    if width > 0:
         ax.annotate(f'{width:0.0f}', xy=(left+width/2, bottom+height/2), ha='center', va='center')

Option 2: 'Ward' as the y-axis

• Use pandas.DataFrame.T to swap the Index and Columns
  • 'Ward' will now be the index and 'Party' will be the columns

# transpose df from the OP so Party is the in the columns and Ward is the index
dft = df.T

# display(dft)
                               Labour  LD  Indep  Green  Tory
Abbey                              16  56      4      0    24
Albrighton                          0   0     32      0    68
Alveley & Claverley                 0  25      0      0    75
Bagley                             30  30      0      0    40
Battlefield                        34   0      0      9    57
Bayston Hill, Column & Sutton      53   4      3      7    33
Belle Vue                          43  28      0      5    24

Using matplotlib from version 3.4.2

# plot
ax = df.T.plot.barh(stacked=True, figsize=(16, 10))

plt.legend(loc='center left', bbox_to_anchor=(1.0, 0.5))

# annotations:
for c in ax.containers:
    
    # format the number of decimal places and replace 0 with an empty string
    labels = [f'{w:.0f}' if (w := v.get_width()) > 0 else '' for v in c ]
    
    ax.bar_label(c, labels=labels, label_type='center')

Using matplotlib before version 3.4.2

# plot
ax = dft.plot.barh(stacked=True, figsize=(16, 10))

plt.legend(loc='center left', bbox_to_anchor=(1.0, 0.5))

# annotations:
for p in ax.patches:
    left, bottom, width, height = p.get_bbox().bounds
    if width > 0:
         ax.annotate(f'{width:0.0f}', xy=(left+width/2, bottom+height/2), ha='center', va='center')