基于第三变量带颜色条的双变量散点图绘制报错及方案咨询

Let's break down what's going wrong with your code first, then walk through two solid solutions to get the exact plot you want.

What Caused the TypeError? #

Your original code has two main issues:

• When you called plt.scatter(df['A'], df['B'], cmap="jet"), you didn't pass the c parameter to link the scatter points to your z (df['C']) values. Without this, the scatter plot doesn't have a data array associated with its colormap, hence the "must first set_array for mappable" error when trying to add a colorbar.
• You're mixing matplotlib.pyplot.scatter and seaborn.lmplot in the same script—these create separate figure objects, so they won't combine into one plot like you're expecting.

Solution 1: Pure Matplotlib Implementation (Simple & Direct) #

This approach uses only matplotlib to create your scatter plot with the Jet colormap, linked directly to your z values:

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

# Generate your data
x = np.random.randint(0, 20, 30)
y = np.random.randint(-5, 5, 30)
z = np.random.randint(-2, 10, 30)
df = pd.DataFrame(data={'A': x, 'B': y, 'C': z})

# Create the scatter plot, linking color to df['C'] with Jet colormap
scatter = plt.scatter(df['A'], df['B'], c=df['C'], cmap='jet')

# Add colorbar with a label for clarity
plt.colorbar(scatter, label='Z Value')

# Add axis labels
plt.xlabel('X (A)')
plt.ylabel('Y (B)')

plt.show()

Key Fixes Here: #

• Added c=df['C'] to plt.scatter to map point colors to your z variable.
• Removed the unnecessary cm.ScalarMappable line—matplotlib handles the colormap mapping automatically when you pass c and cmap.
• No conflicting seaborn calls, so everything stays in one figure.

Solution 2: Seaborn with Jet Colormap #

If you prefer using seaborn, you don't have to abandon the Jet colormap. You can explicitly set palette='jet' in seaborn.scatterplot (a better fit than lmplot for pure scatter plots):

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

# Generate data (same as before)
x = np.random.randint(0, 20, 30)
y = np.random.randint(-5, 5, 30)
z = np.random.randint(-2, 10, 30)
df = pd.DataFrame(data={'A': x, 'B': y, 'C': z})

# Create seaborn scatter plot with Jet colormap
sns.scatterplot(data=df, x='A', y='B', hue='C', palette='jet', s=50)

# Customize the colorbar label
plt.colorbar(label='Z Value')

# Add axis labels
plt.xlabel('X (A)')
plt.ylabel('Y (B)')

plt.show()

Why This Works: #

• sns.scatterplot supports the palette parameter, so you can directly specify 'jet' to use that colormap.
• Using hue='C' links the point colors to your z values, just like the c parameter in matplotlib.
• We use scatterplot instead of lmplot because lmplot is designed for regression plots (even with fit_reg=False, it adds extra overhead and creates a FacetGrid, which is unnecessary here).

Both solutions will produce a scatter plot where points are colored based on your z values using the Jet colormap, matching the style of your reference example.

内容的提问来源于stack exchange，提问作者gis20