如何在Python Plotly的px.imshow热力图中实现散点抖动

我用Plotly制作了风险矩阵热力图，叠加散点表示单个项目，但同一单元格（比如影响中等、可能性低的单元格）有多个项目时，散点会重叠。怎么给散点加抖动效果让它们在单元格内不重叠？示例代码如下：

import plotly.express as px
import plotly.graph_objects as go

fig = px.imshow([[3, 5, 5],
                 [1, 3, 5],
                 [1, 1, 3]],
                color_continuous_scale='Reds',
                labels=dict(x="Likelihood", y="Impact"),
                x=['Low', 'Medium', 'High'],
                y=['High', 'Medium', 'Low']
                )
fig.add_trace(go.Scatter(x=["Low"], y=["Medium"],
                         name="project1",
                         marker=dict(color='black', size=16)))
fig.show()

要实现这个效果，核心思路是给同一单元格内的散点坐标添加微小的随机偏移量（抖动）。Plotly的类别轴底层是用连续数值（0、1、2...）映射类别的，基于这些数值生成偏移，既能保证点在单元格内，又能避免重叠。

具体实现代码 #

import plotly.express as px
import plotly.graph_objects as go
import numpy as np

# 定义风险矩阵基础数据
heatmap_data = [[3, 5, 5],
                [1, 3, 5],
                [1, 1, 3]]
x_labels = ['Low', 'Medium', 'High']
y_labels = ['High', 'Medium', 'Low']

# 生成热力图
fig = px.imshow(heatmap_data,
                color_continuous_scale='Reds',
                labels=dict(x="Likelihood", y="Impact"),
                x=x_labels,
                y=y_labels)

# 模拟多项目数据，包含同一单元格的项目
projects = [
    {"name": "project1", "likelihood": "Low", "impact": "Medium"},
    {"name": "project2", "likelihood": "Low", "impact": "Medium"},
    {"name": "project3", "likelihood": "Medium", "impact": "High"},
    {"name": "project4", "likelihood": "Medium", "impact": "High"}
]

# 映射类别标签到数值索引
x_indices = {label: idx for idx, label in enumerate(x_labels)}
y_indices = {label: idx for idx, label in enumerate(y_labels)}

jitter_amount = 0.2  # 抖动幅度，建议不超过0.5（避免跨单元格）
x_jittered = []
y_jittered = []
proj_names = []

# 为每个项目生成带抖动的坐标
for proj in projects:
    x_idx = x_indices[proj["likelihood"]]
    y_idx = y_indices[proj["impact"]]
    # 添加随机偏移
    x_jittered.append(x_idx + np.random.uniform(-jitter_amount, jitter_amount))
    y_jittered.append(y_idx + np.random.uniform(-jitter_amount, jitter_amount))
    proj_names.append(proj["name"])

# 添加带抖动的散点，可选显示项目名称
fig.add_trace(go.Scatter(
    x=x_jittered,
    y=y_jittered,
    mode='markers+text',
    text=proj_names,
    textposition='top center',
    name="Projects",
    marker=dict(color='black', size=16)
))

# 保持轴的类别标签显示，不切换为数值
fig.update_layout(
    xaxis=dict(tickmode='array', tickvals=list(x_indices.values()), ticktext=list(x_indices.keys())),
    yaxis=dict(tickmode='array', tickvals=list(y_indices.values()), ticktext=list(y_indices.keys()))
)

fig.show()

关键说明 #

• jitter_amount：控制抖动幅度，值越大点分散越开，但不要超过0.5，否则点会跑到相邻单元格
• 如果需要规整排列而非随机抖动，可以预先设置固定偏移量（比如同一单元格内的点按(±0.1, ±0.1)排列），适合项目数量固定的场景
• 启用mode='markers+text'可以同时显示项目名称，方便快速识别每个散点对应的项目

内容的提问来源于stack exchange，提问作者Subaru Spirit