如何优化Seaborn lineplot的数值标注效果？

数据集 #

初始实现代码 #

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
sns.set_style('darkgrid')

fig, ax = plt.subplots(figsize=(10, 6))
sns.lineplot(x=df['month'], y=df['mentor_cnt'], color = 'b', label='Mentors', markers=True, marker='o')
sns.lineplot(x=df['month'], y=df['mentee_cnt'], color = 'r', label='Mentee', markers=True, marker='o')
plt.ylim(min(df['mentee_cnt'])-60, max(df['mentee_cnt'])+50)
columns =['mentor_cnt', 'mentee_cnt']
colors = ['b', 'w']
facecolors = ['none', 'r']
x_position = [0, -10]
y_position = [-18, 5]

for col, fc, c, x_p, y_p in zip(columns, facecolors, colors, x_position, y_position):
    for x, y in zip(df['month'], df[col]):
        label = '{:.0f}'.format(y)
        plt.annotate(
            label,
            (x, y),
            textcoords='offset points',
            xytext=(x_p, y_p),
            ha='center',
            color=c
        ).set_bbox(dict(facecolor=fc, alpha=0.5, boxstyle='round', edgecolor='none'))
ax.legend()
plt.xlabel('Month')
plt.ylabel('Number of persons')
plt.title('Mentee, mentor')
plt.show()

初始效果 #

导师（Mentors）的折线标注效果尚可，但学员（Mentee）的数值标注拥挤重叠，可读性差，尝试用set_bbox提升效果也不佳。

优化思路 #

通过仅标注关键分位数和末尾数值减少标注数量，避免拥挤，同时调整标注位置参数提升可读性。

优化后代码 #

fig, ax = plt.subplots(figsize=(10, 6))
sns.lineplot(x=df['month'], y=df['mentor_cnt'], color='b', label='Mentors', markers=True, marker='o')
sns.lineplot(x=df['month'], y=df['mentee_cnt'], color='r', label='Mentee', markers=True, marker='o')
plt.ylim(min(df['mentee_cnt'])-60, max(df['mentee_cnt'])+50)
columns = ['mentor_cnt', 'mentee_cnt']
colors = ['b', 'r']
facecolors = ['none', 'r']
x_position = [0, -15]
y_position = [-18, 3]

for col, fc, c, x_p, y_p in zip(columns, facecolors, colors, x_position, y_position):
    for x, y in zip(df['month'], df[col]):
        # 生成包含分位数和末尾值的关键数值列表
        lst = [0, 0.25, 0.5, 0.75, 1]
        describe_nearest = []
        [describe_nearest.append(df[col].quantile(el, interpolation='nearest')) for el in lst]
        describe_nearest.append(df[col].values[-1::][0])
        # 仅对关键数值添加标注
        if y in describe_nearest:
            label = '{:.0f}'.format(y)
            plt.annotate(
                label,
                (x, y),
                textcoords='offset points',
                xytext=(x_p, y_p),
                ha='center',
                color=c
            )
ax.legend()
plt.xlabel('Month')
plt.ylabel('Number of persons')
plt.title('Mentee, mentor')
plt.show()

优化后效果 #

仅保留关键分位数（0、25%、50%、75%、100%分位）和最后一个数值的标注，标注布局清晰，学员折线的标注可读性显著提升。

内容的提问来源于stack exchange，提问作者John Doe