如何从CSV文件的Pandas DataFrame中查找局部最大值并标注到图表？

1. 数据准备（读取CSV或导入已有数据） #

如果是从CSV读取数据，直接用pd.read_csv()即可；针对你提供的示例数据，可按如下方式构建DataFrame：

import pandas as pd
import matplotlib.pyplot as plt
from scipy.signal import find_peaks

# 你的x、y数据
x = [1598.78, 1596.85, 1594.92, 1592.99, 1591.07, 1589.14, 1587.21, 1585.28, 1583.35, 1581.42, 1579.49, 1577.57, 1575.64, 1573.71, 1571.78, 1569.85, 1567.92, 1565.99, 1564.07, 1562.14, 1560.21, 1558.28, 1556.35, 1554.42, 1552.49, 1550.57, 1548.64, 1546.71, 1544.78, 1542.85, 1540.92, 1538.99, 1537.07, 1535.14, 1533.21, 1531.28, 1529.35, 1527.42, 1525.49, 1523.57, 1521.64, 1519.71, 1517.78, 1515.85, 1513.92, 1511.99, 1510.07, 1508.14, 1506.21, 1504.28, 1502.35, 1500.42, 1498.49, 1496.57, 1494.64, 1492.71, 1490.78, 1488.85, 1486.92, 1484.99, 1483.07, 1481.14, 1479.21, 1477.28, 1475.35, 1473.42, 1471.49, 1469.57, 1467.64, 1465.71, 1463.78, 1461.85, 1459.92, 1457.99, 1456.07, 1454.14, 1452.21, 1450.28, 1448.35, 1446.42, 1444.49, 1442.57, 1440.64, 1438.71, 1436.78, 1434.85, 1432.92, 1430.99, 1429.07, 1427.14, 1425.21, 1423.28, 1421.35, 1419.42, 1417.49, 1415.57, 1413.64, 1411.71, 1409.78, 1407.85, 1405.92, 1403.99, 1402.07, 1400.14]
y = [0.640, 0.624, 0.609, 0.594, 0.581, 0.569, 0.558, 0.547, 0.537, 0.530, 0.523, 0.516, 0.508, 0.502, 0.497, 0.491, 0.487, 0.484, 0.481, 0.480, 0.479, 0.482, 0.490, 0.503, 0.520, 0.542, 0.566, 0.586, 0.600, 0.606, 0.593, 0.569, 0.557, 0.548, 0.538, 0.531, 0.527, 0.524, 0.522, 0.522, 0.523, 0.525, 0.526, 0.527, 0.530, 0.534, 0.536, 0.539, 0.547, 0.553, 0.557, 0.563, 0.573, 0.599, 0.654, 0.738, 0.852, 0.891, 0.810, 0.744, 0.711, 0.694, 0.686, 0.683, 0.683, 0.690, 0.700, 0.706, 0.713, 0.723, 0.731, 0.732, 0.737, 0.756, 0.779, 0.786, 0.790, 0.794, 0.802, 0.815, 0.827, 0.832, 0.831, 0.826, 0.823, 0.828, 0.834, 0.834, 0.832, 0.832, 0.831, 0.825, 0.816, 0.804, 0.798, 0.794, 0.786, 0.775, 0.764, 0.752, 0.739, 0.722, 0.708, 0.697]

df = pd.DataFrame({"x": x, "y": y})

2. 检测局部最大值 #

用scipy.signal.find_peaks精准检测峰值，可通过参数过滤无效小峰：

# 查找峰值：height设置最小高度阈值，distance设置峰值间最小数据点间隔
peaks, props = find_peaks(df["y"], height=0.5, distance=10)

# 提取峰值对应的x、y数据
peak_points = df.iloc[peaks]

3. 绘制曲线并标注峰值 #

用Matplotlib绘制原始曲线，再通过plt.text在峰值位置标注数值：

plt.figure(figsize=(12, 6))
plt.plot(df["x"], df["y"], color="#1f77b4", linewidth=1.5, label="数据曲线")

# 逐个标注峰值的y值，调整偏移量避免与曲线重叠
for _, peak in peak_points.iterrows():
    plt.text(peak["x"], peak["y"] + 0.015, 
             f"{peak['y']:.3f}", 
             ha="center", va="bottom", 
             fontsize=10, color="#ff7f0e")

# 图表基础设置
plt.xlabel("X轴", fontsize=12)
plt.ylabel("Y轴", fontsize=12)
plt.title("带局部最大值标注的曲线", fontsize=14)
plt.legend()
plt.grid(alpha=0.3)
plt.show()

参数调整说明 #

• find_peaks的height：过滤低于该值的小波动，避免误判；
• find_peaks的distance：控制两个峰值的最小间隔，避免相邻近峰被识别；
• plt.text的偏移量：通过peak["y"] + 0.015调整文字在峰值上方的位置，可根据图表比例修改数值。

内容的提问来源于stack exchange，提问作者cbornes