在Pandas中绘制多层索引DataFrame：按symbol分线同图展示

Got it, let's walk through how to plot each symbol as a distinct line on the same chart, assuming your multi-index DataFrame uses something like (date/time, symbol) as its index (super common for time-series data with multiple assets). I'll cover two straightforward methods, plus some tips for customization.

First, Let's Set Up a Sample DataFrame (to match your structure) #

If you don't already have one, here's a quick example to work with—this mimics the kind of data you're describing:

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

# Create sample multi-index data
dates = pd.date_range('2024-01-01', periods=15)
symbols = ['AAPL', 'MSFT', 'GOOGL']
multi_idx = pd.MultiIndex.from_product([dates, symbols], names=['date', 'symbol'])
# Generate cumulative random values to simulate trends
values = np.random.randn(45).cumsum()
df = pd.DataFrame({'value': values}, index=multi_idx)

Method 1: Unstack the Symbol Index (Simplest Approach) #

The easiest way is to "unstack" the symbol level from your index into columns. Pandas' built-in plot() will automatically draw a line for each column:

# Unstack the 'symbol' index level to turn it into columns
df_unstacked = df.unstack(level='symbol')
# Clean up the column names (unstack creates a multi-level column by default)
df_unstacked.columns = df_unstacked.columns.droplevel(0)

# Plot all symbols on the same chart
df_unstacked.plot(figsize=(10, 6), linewidth=2)
plt.title('Symbol Value Trends Over Time')
plt.xlabel('Date')
plt.ylabel('Value')
plt.legend(title='Ticker Symbol')
plt.grid(alpha=0.3)
plt.show()

This works because unstacking transforms your data into a wide format where each symbol is a separate column—exactly what Pandas needs to plot individual lines.

Method 2: Group by Symbol (More Flexible for Customization) #

If you want more control over each line's style (colors, markers, etc.), use groupby() to iterate over each symbol and plot manually:

plt.figure(figsize=(10, 6))

# Group the DataFrame by the 'symbol' index level
for symbol, group_data in df.groupby(level='symbol'):
    # Drop the symbol index level so we can use date as the x-axis
    group_data = group_data.reset_index(level='symbol', drop=True)
    # Plot with custom styling (adjust as needed)
    group_data.plot(
        label=symbol,
        linewidth=2,
        marker='o',
        markersize=4,
        legend=True
    )

plt.title('Customized Symbol Value Trends')
plt.xlabel('Date')
plt.ylabel('Value')
plt.legend(title='Ticker Symbol')
plt.grid(alpha=0.3)
plt.show()

This method lets you tweak each line individually—great if you want to highlight specific symbols with unique colors or markers.

Quick Tips #

• If your symbol is in a different index level (e.g., first instead of second), adjust the level parameter in unstack() or groupby() (use level=0 instead of level='symbol').
• If your DataFrame has multiple value columns, specify which one to plot: df_unstacked['your_value_column'].plot().
• Use Pandas' plot styling options (like color, linestyle, marker) to make your chart easier to read.

内容的提问来源于stack exchange，提问作者Evan Kim