如何绘制分类数据并通过颜色、点大小编码？求技术术语与工具

Hey there! Let's tackle your two questions about categorical data visualization—including the key terminology you're curious about, plus practical code examples with different tools.

First, let's cover the core terminology here:

• This type of visualization is a variant of a scatter plot, where we use visual channel encoding to add extra dimensions to the plot.
• Using color for data mapping is called hue encoding, and using point size is size encoding—both are effective ways to represent additional categorical or numerical attributes alongside your main x/y axes.

Here are practical implementations with common tools:

Using Seaborn (Matplotlib-based) #

Seaborn makes categorical visualization straightforward with built-in parameters:

import seaborn as sns
import pandas as pd

# Sample categorical dataset
df = pd.DataFrame({
    "Category": ["A", "A", "B", "B", "C", "C"],
    "X_Value": [1, 2, 3, 4, 5, 6],
    "Y_Value": [5, 7, 3, 8, 2, 6],
    "Group": ["Control", "Treatment", "Control", "Treatment", "Control", "Treatment"],  # Color mapping
    "Metric": [10, 20, 15, 25, 8, 18]  # Size mapping
})

# Plot with hue (color) and size encoding
sns.scatterplot(data=df, x="X_Value", y="Y_Value", hue="Group", size="Metric")

Using Plotly (Interactive) #

For interactive plots that let users hover to see details:

import plotly.express as px

fig = px.scatter(df, x="X_Value", y="Y_Value", color="Group", size="Metric", size_max=60)
fig.show()

For this specific requirement, the technical term you're looking for is a categorical scatter plot (or an extended dot plot)—since both your x and y axes are categorical variables, and you're using visual channels to encode two more attributes.

Below are implementations with your mentioned tool (pdvega) plus alternatives:

Using pdvega #

import pandas as pd
import pdvega

# Sample dataset matching your requirements
df = pd.DataFrame({
    "Participant": ["P1", "P1", "P2", "P2", "P3", "P3"],
    "Skill": ["Python", "SQL", "Python", "R", "SQL", "R"],
    "Proficiency": ["High", "Medium", "Medium", "High", "Low", "Medium"],  # Color mapping
    "Years_Experience": [5, 3, 2, 6, 1, 4]  # Size mapping
})

# Create the plot with pdvega
df.vgplot.scatter(
    x="Participant",
    y="Skill",
    color="Proficiency",
    size="Years_Experience",
    size_scale={"range": [50, 300]}  # Adjust size range for visibility
)

Using Seaborn (Static, Customizable) #

import seaborn as sns
import matplotlib.pyplot as plt

plt.figure(figsize=(8, 5))
sns.scatterplot(
    data=df,
    x="Participant",
    y="Skill",
    hue="Proficiency",
    size="Years_Experience",
    sizes=(50, 300),  # Control point size range
    alpha=0.8  # Add transparency for overlapping points
)
plt.title("Participant Skill Matrix with Proficiency & Experience")
plt.show()

Key Terminology Recap #

• Visual Channels: The visual properties used to represent data (here: hue/color and size)
• Categorical Scatter Plot: A scatter plot where one or both axes are categorical variables, ideal for showing relationships between categorical attributes
• Data Mapping: The process of linking dataset columns (e.g., Proficiency, Years_Experience) to visual elements (color, size)

内容的提问来源于stack exchange，提问作者ralien