A bubble plot is a scatterplot with a third dimension: the size of the markers. It is even possible to add a fourth dimension using colors. It can be done using the cmap argument of the scatter() function of matplotlib.

Libraries & Dataset

We will start by importing the necessary libraries and loading the dataset.

Since bubble plots requires numerical values, we need to have quantitative data in our dataset.

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

# create data
df = pd.DataFrame({
      'x': np.random.rand(40),
      'y': np.random.rand(40),
      'z': np.random.rand(40),
      'cat': np.random.choice(['A', 'B', 'C'], 40)
   })
df.head()
xyzcat
00.5816790.2399450.266493A
10.7762070.7348590.788658B
20.6723330.0554080.865137C
30.1512270.8203740.186279A
40.5171760.4036050.312726A

Map with numerical values

First, we initiate a simple bubble plot using the scatter() function from matplotlib. By incorporating the cmap argument, we introduce a fourth dimension to our plot.

Here, the y variable dictates the color of the markers—observed as progressively darker hues as values ascend on the y-axis.

Map a color to a bubble plot: Map with numerical values

# Initiate the chart
fig, ax = plt.subplots(figsize=(8,6), dpi=200)

# We map the color to the X axis value.
ax.scatter(
   x=df['x'],
   y=df['y'],
   s=df['z']*2000,
   c=df['y'],
   cmap="Blues",
   alpha=0.4,
   edgecolors="grey",
   linewidth=2
)

# Add titles (main and on axis)
ax.set_xlabel("the X axis")
ax.set_ylabel("the Y axis")
ax.set_title("A colored bubble plot")

# Show the graph
plt.show()

Map with categorical values with cmap

To create a bubble plot with categorical color mapping and a legend in Matplotlib:

  1. Encode Categories: Convert category names into numerical codes using pd.factorize.

  2. Initialize Chart: Set up the plotting area using fig, ax = plt.subplots.

  3. Create Bubble Plot: Plot the data points using ax.scatter. The x and y parameters determine the position of the bubbles, s controls their size, and c sets their color based on category codes. cmap defines the color map, and alpha sets the transparency.

  4. Add Legend:

    • Import Patch from matplotlib.patches.
    • Create legend handles by iterating over the unique category names and their corresponding codes. Each handle is a Patch object with the color from the scatter plot's colormap.
    • Add the legend to the plot with ax.legend, setting the handles, title, and location.

  5. Customize Axes and Display: Label the axes and title, then show the plot with plt.show().

Map a color to a bubble plot: Map with categorical values with cmap

# Encode the categories as integers
category_codes, category_names = pd.factorize(df['cat'])

# Initiate the chart
fig, ax = plt.subplots(figsize=(8,6), dpi=200)

# Create the bubble plot
scatter = ax.scatter(
   x=df['x'],
   y=df['y'],
   s=df['z']*500,
   c=category_codes,
   cmap='viridis',
   alpha=0.6
)

# Legend
from matplotlib.patches import Patch
handles = [Patch(facecolor=scatter.cmap(scatter.norm(i)), label=label)
           for i, label in enumerate(category_names)]
ax.legend(handles=handles, title='Categories', loc='upper right')

# Custom axes and display chart
ax.set_xlabel('X-axis')
ax.set_ylabel('Y-axis')
ax.set_title('Bubble plot with categorical color mapping')
plt.show()

Map with categorical values with specific colors

If you want to use your own colors for the categories, you can create a dictionary that maps category names to colors. Then, you can pass the colors to the c parameter of the scatter() function.

However, there is no know method for avoiding the for loop when creating the scatter plot.

Map a color to a bubble plot: Map with categorical values with specific colors

# Initiate the chart
fig, ax = plt.subplots(figsize=(8,6), dpi=200)

# Define color mapping
colors = {'A':'darkred', 'B':'skyblue', 'C':'green'}

# Iterate over each group value
for group, color in colors.items():
    ax.scatter(
        x=df['x'][df['cat']==group],
        y=df['y'][df['cat']==group],
        s=df['z'][df['cat']==group]*1000,
        c=color,
        alpha=0.4
    )

# Add titles (main and on axis)
ax.set_xlabel("the X axis")
ax.set_ylabel("the Y axis")
plt.title("A colored bubble plot")

# Show the graph
plt.show()

Going further

You might be interested in: