如何为条形图设置文字注释的水平对齐方式

Question

我用这段代码绘制了一些数据：

plt.rcdefaults()
fig, ax = plt.subplots(figsize=(10,6))
# Example data
people = data_year[data_year.columns[1]]
y_pos = np.arange(len(people))
#performance = 3 + 10 * np.random.rand(len(people))
performance=data_year[data_year.columns[2]]
error = np.random.rand(len(people))
rects=ax.barh(y_pos, performance, xerr=error, align='center')
ax.invert_yaxis()  # labels read top-to-bottom
ax.set_xlabel('population (thousands)' , fontsize=15)
ax.text(0.9, 0.1, year,verticalalignment='bottom', horizontalalignment='right',transform=ax.transAxes,color='red',size=30)
for i, (p, pr) in enumerate(zip(data_year["Country or area"], data_year["Urban Agglomeration"])):
    plt.text(s=p, x=1, y=i, color='black', verticalalignment="top", horizontalalignment="left", size=8)
    plt.text(s=pr , x=1, y=i, color='black',verticalalignment="bottom",horizontalalignment="left", size=10)
display(plt.show)

结果是这样的 但我需要的结果是这样的

如何更改标签的位置

Answer 1

• 使用.patches提取的矩形尺寸来设置x位置，也可以使用ha='right'。
  • 另见SO solution
  • matplotlib: Text properties and layout
• 来自data.un.org的数据，为2015年和所有国家选择

从下载的文件中选择数据

• 从链接下载数据文件，或使用答案底部的示例数据框df

import pandas as pd
import matplotlib.pyplot as plt

# load the files
data = pd.read_csv('data/UNdata_Export_20210131_205829459.csv')

# filter data
data = data.loc[(data['City type'] == 'City proper') & (data['Sex'] == 'Both Sexes')][['Country or Area', 'City', 'Value', 'Year']]

# sort data
data = data.sort_values('Value', ascending=False)

# select top 10 rows
df = data.iloc[:10]

用OP法绘图

• 此实现不必要地将数据帧列提取到新变量中（例如people）

people = df[df.columns[2]]
y_pos = np.arange(len(people))

error = np.random.rand(len(people))

fig, ax = plt.subplots(figsize=(10,6))

rects = ax.barh(y_pos, people, xerr=error, align='center')

ax.invert_yaxis()  # labels read top-to-bottom
ax.set_xlabel('population' , fontsize=15)

ax.text(0.9, 0.1, 2015, va='bottom', ha='right', transform=ax.transAxes, color='red',size=30)

for i, (p, pr) in enumerate(zip(df['Country or Area'], df['City'])):

    # extract the rectangles from ax
    rectangle = ax.patches[i]
    
    # get the x for the bar
    x = rectangle.get_x()
    
    # get the width
    width = rectangle.get_width()
    
    # calculate the label position
    x_label = x + width
    
    # use the rectangle x to plot the text and set the ha to right instead of left
    plt.text(s=p, x=x_label, y=i, color='black', va="top", ha="right", size=8, weight='bold')
    plt.text(s=pr, x=x_label, y=i, color='black', va="bottom", ha="right", size=10, weight='bold')
    
display(plt.show)

简化实施

• 绘制特定的dataframe列，而不将其提取到新变量
• 用^{}代替^{}

# plot the dataframe
ax = df.plot.barh(y='Value', figsize=(10, 6), legend=False, width=0.8)
ax.set_xlabel('population' , fontsize=15)

ax.set(yticklabels=[])  # remove the y tick labels
ax.tick_params(left=False)  # remove the ticks

ax.invert_yaxis()

ax.text(0.9, 0.1, 2015, va='bottom', ha='right', transform=ax.transAxes, color='red',size=30)

for i, (p, pr) in enumerate(zip(df['Country or Area'], df['City'])):

    # extract the rectangles from ax
    rectangle = ax.patches[i]
    
    # get the x for the bar
    x = rectangle.get_x()
    
    # get the width
    width = rectangle.get_width()
    
    # calculate the label position
    x_label = x + width
    
    # use the rectangle x to plot the text and set the ha to right instead of left
    plt.text(s=p, x=x_label, y=i, color='black', va="top", ha="right", size=8, weight='bold')
    plt.text(s=pr, x=x_label, y=i, color='black', va="bottom", ha="right", size=10, weight='bold')
    
plt.show()

绘图结果

df样本

• 如果网站数据不再可用，则测试数据

data = {'Country or Area': ['Republic of Korea', 'Japan', 'Mexico', 'United States of America', 'China, Hong Kong SAR', 'Singapore', 'Chile', 'Australia', 'Australia', 'Jordan'], 'City': ['SEOUL', 'TOKYO', 'MEXICO, CIUDAD DE', 'New York (NY)', 'HONG KONG SAR', 'SINGAPORE', 'SANTIAGO', 'Sydney', 'Melbourne', 'AMMAN'], 'Value': [9860372.0, 9272740.0, 8854560.0, 8550405.0, 7291300.0, 5535002.0, 5150010.0, 4526479.0, 4353514.0, 4007526.0], 'Year': ['2015', '2015', '2015', '2015', '2015', '2015', '2015', '2015', '2015', '2015']}
df = pd.DataFrame(data)

display(df)
            Country or Area               City      Value  Year
0         Republic of Korea              SEOUL  9860372.0  2015
1                     Japan              TOKYO  9272740.0  2015
2                    Mexico  MEXICO, CIUDAD DE  8854560.0  2015
3  United States of America      New York (NY)  8550405.0  2015
4      China, Hong Kong SAR      HONG KONG SAR  7291300.0  2015
5                 Singapore          SINGAPORE  5535002.0  2015
6                     Chile           SANTIAGO  5150010.0  2015
7                 Australia             Sydney  4526479.0  2015
8                 Australia          Melbourne  4353514.0  2015
9                    Jordan              AMMAN  4007526.0  2015