Python matplotlib绘制分季节周内逐时均值连续时序图

你当前代码无法实现折线连续排布的核心原因是：所有星期几的折线都使用0-23小时作为x轴坐标，导致7条线完全重叠在单日24小时区间内。只需为不同星期的小时值增加对应天数的偏移量，构造总长度168小时（7天*24小时）的连续x轴，再调整刻度标签即可匹配目标效果。

修改后可直接运行的完整代码 #

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib as mpl
import numpy as np

# ---------------------- 1. 构造测试数据（和原逻辑一致） ----------------------
df = pd.DataFrame()
df["date"] = pd.date_range(start='2018', end='2019', freq = "15Min")
df=df.set_index(["date"])

dw_mapping={0: 'Mon', 1: 'Tue', 2: 'Wed', 3: 'Thu', 4: 'Fri', 5: 'Sat', 6: 'Sun'} 
df['Day']=df.index.weekday.map(dw_mapping)

season_mapping={1: 'Winter',
            2: 'Winter',
            3: 'eS/lF',
            4: 'eS/lF',
            5: 'lS/eF',
            6: 'Summer',
            7: 'Summer',
            8: 'Summer',
            9: 'lS/eF',
            10:'eS/lF', 
            11:'eS/lF',
            12:'Winter'}
df['season']=df.index.month.map(season_mapping)

df["B1W"] = 1
df['B1W'] = np.where(df['season'] == 'Winter', df['B1W'] * np.random.randint(30, 60, df.shape[0]), df["B1W"])
df['B1W'] = np.where(df['season'] == 'eS/lF', df['B1W'] *  np.random.randint(20, 50, df.shape[0]), df["B1W"])
df['B1W'] = np.where(df['season'] == 'lS/eF', df['B1W'] *  np.random.randint(10, 30, df.shape[0]), df["B1W"])
df['B1W'] = np.where(df['season'] == 'Summer', df['B1W'] * np.random.randint(0, 10, df.shape[0]), df["B1W"])

# ---------------------- 2. 绘图配置 ----------------------
mpl.rcParams['figure.dpi'] = 100
plt.style.use('ggplot')
ymin, ymax = 0, 60
# 固定星期顺序
day_list = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
# 子图和季节的映射关系
season_ax_map = [
    ("Summer", (0,0)),
    ("eS/lF", (0,1)),
    ("lS/eF", (1,0)),
    ("Winter", (1,1))
]

fig, axes = plt.subplots(2, 2)
plt.subplots_adjust(wspace=0.2, hspace=0.3, bottom=0.2)
fig.suptitle("Building 1", fontsize=16)
fig.set_figheight(10)
fig.set_figwidth(15)

# ---------------------- 3. 分季节绘制连续折线 ----------------------
for season, (ax_row, ax_col) in season_ax_map:
    ax = axes[ax_row, ax_col]
    # 筛选当前季节数据，固定星期字段顺序
    season_data = df.loc[df["season"]==season].copy()
    season_data['Day'] = pd.Categorical(season_data['Day'], categories=day_list, ordered=True)
    # 计算分星期、分小时的均值
    hourly_avg = season_data.groupby([season_data.index.hour, 'Day'])["B1W"].mean().unstack()
    
    # 为每个星期的小时值加24*天数偏移，构造连续x轴
    for day_idx, day_name in enumerate(day_list):
        x = np.arange(24) + day_idx * 24
        y = hourly_avg[day_name].values
        ax.plot(x, y, label=day_name)
    
    # 坐标轴配置
    ax.set_ylim(ymin, ymax)
    ax.set_title(season)
    ax.set_ylabel('Power in MW')
    # x轴刻度对应每天起始位置，标签为星期名
    ax.set_xticks(np.arange(0, 7*24, 24))
    ax.set_xticklabels(day_list)
    ax.set_xlabel('Time')

# 统一配置底部图例
handles, labels = axes[1,1].get_legend_handles_labels()
axes[1,1].legend(handles, labels, loc=1, ncol=7, bbox_to_anchor=(0.638,-0.2), frameon=True)

plt.show()

关键改动说明 #

• 提前固定周一到周日的排序规则，避免分组聚合后星期顺序错乱
• 不再将0-23小时作为统一x轴，而是给每个星期的小时值增加星期序号*24的偏移量，构造总长度168小时的连续时间轴，实现7天折线首尾相接
• 调整x轴刻度位置为每天的0点位置（0、24、48...144），刻度标签替换为对应星期名，和目标效果图的坐标轴样式完全匹配
• 保留了原有的子图布局、配色风格、标题与图例位置，无需调整原有可视化配置习惯

目标效果参考：

原实现效果参考：

内容的提问来源于stack exchange，提问作者Freedo