关于Seaborn直方图标签错误及展示样式调整的技术咨询

Hey there! Let's work through fixing your Seaborn plot to get exactly what you want—proper labels and line-only styling for both Bot and Human accounts.

First, let's break down the issues you're facing:

• The legend has mismatched markers (a box and a line) instead of consistent line markers
• You want both groups to display as lines (no filled areas) with distinct colors
• The label alignment might be off if the order doesn't match your data's bot column values

Here's the revised code with explanations for each change:

import matplotlib.pyplot as plt
from matplotlib.lines import Line2D
import seaborn as sns

plt.figure(figsize=(7,5), dpi=100)

# Core plot setup: Remove fill to show only lines, disable auto-legend
sns.histplot(
    x='hour_created', 
    data=training1_data, 
    hue='bot', 
    kde=True, 
    stat='density', 
    common_bins=True, 
    element='step', 
    fill=False,  # 关键改动：去掉填充区域，只保留线条
    legend=False  # 关闭Seaborn自动图例，避免标记冲突
)

# Create custom legend with line markers (instead of boxes)
# Use Seaborn's default color palette to match plot colors
custom_legend = [
    Line2D([0], [0], color=sns.color_palette()[0], lw=2),
    Line2D([0], [0], color=sns.color_palette()[1], lw=2)
]
# Adjust labels here to match your data's bot category order!
# Run `print(training1_data['bot'].unique())` to check the order of values
plt.legend(
    custom_legend, 
    ['Bot', 'Human'], 
    title='Account Type', 
    loc='best'
)

# Keep your original plot formatting
plt.title("Time of Day Account was Created", fontsize=12)
plt.xlabel('Time of Day (24h)', fontsize=10)
plt.ylabel('Density', fontsize=10)
plt.xlim(0, 24)
plt.ylim(0, 0.05)
sns.despine()

# plt.savefig('imgs/hour_created.png');

Key fixes explained:

1. Line-only display: Adding fill=False removes the colored fill under the step histogram, so both the histogram and KDE lines are visible without any filled boxes—exactly what you asked for.
2. Consistent legend markers: By creating custom Line2D objects for the legend, we replace the default box markers with line markers, so both entries in the legend match the plot's line style.
3. Label alignment: Double-check the order of ['Bot', 'Human'] against your data's bot column. If you run print(training1_data['bot'].unique()), you'll see the order Seaborn uses for hue—adjust the labels in plt.legend() if they're reversed.

That should get your plot looking exactly how you want it!

备注：内容来源于stack exchange，提问作者WT YXK