如何在Plotly销售漏斗中显示基于网站访客的百分比？

Plotly默认的percent initial是基于整个漏斗的第一个系列（所有月份访客总和），而非单月自身的访客数；percent previous则是相对上一步骤的占比，都无法满足你需要的“各步骤百分比相对于当月网站访客”的需求。要实现该效果，需手动计算百分比并自定义漏斗的显示文本。

实现步骤 #

1. 为数据框添加各步骤相对当月访客的百分比列，精确到指定小数位数
2. 在每个go.Funnel组件中，通过text参数自定义显示内容（数值+自定义百分比），并将textinfo设为'text'以启用自定义文本

修改后的完整代码 #

import plotly.graph_objs as go
import pandas as pd

# 初始化示例数据（如果已有merged_month_df可跳过此部分）
data = [
    ["2022-09",778,186,40,2],
    ["2022-10",843,219,68,21],
    ["2022-11",786,195,66,43],
    ["2022-12",805,200,65,47],
    ["2023-01",852,215,77,63],
    ["2023-02",32,9,3,3]
]
merged_month_df = pd.DataFrame(data, columns=["month","visit_site_month","reg_month","application","game"])

# 计算各步骤相对当月访客的百分比，保留2位小数
merged_month_df['reg_pct'] = (merged_month_df['reg_month'] / merged_month_df['visit_site_month'] * 100).round(2)
merged_month_df['app_pct'] = (merged_month_df['application'] / merged_month_df['visit_site_month'] * 100).round(2)
merged_month_df['game_pct'] = (merged_month_df['game'] / merged_month_df['visit_site_month'] * 100).round(2)

# 创建漏斗步骤
steps = [
    go.Funnel(
        name='Visits',
        x=merged_month_df['visit_site_month'],
        y=merged_month_df['month'],
        textposition='inside',
        textinfo="value+percent initial",
        marker={'color': '#3b78e7'}
    ),
    go.Funnel(
        name='Registrations',
        x=merged_month_df['reg_month'],
        y=merged_month_df['month'],
        textposition='inside',
        textinfo='text',
        text=[f"{val} ({pct}%)" for val, pct in zip(merged_month_df['reg_month'], merged_month_df['reg_pct'])],
        marker={'color': '#44a9ff'}
    ),
    go.Funnel(
        name='Applications',
        x=merged_month_df['application'],
        y=merged_month_df['month'],
        textposition='inside',
        textinfo='text',
        text=[f"{val} ({pct}%)" for val, pct in zip(merged_month_df['application'], merged_month_df['app_pct'])],
        marker={'color': '#88c2ff'}
    ),
    go.Funnel(
        name='Game Plays',
        x=merged_month_df['game'],
        y=merged_month_df['month'],
        textposition='inside',
        textinfo='text',
        text=[f"{val} ({pct}%)" for val, pct in zip(merged_month_df['game'], merged_month_df['game_pct'])],
        marker={'color': '#c7e2ff'}
    )
]

# 布局设置
layout = go.Layout(
    title='Funnel Chart of Sales Dynamics',
    funnelmode='stack',
    showlegend=True,
    legend=dict(x=1.0, y=0.8),
    xaxis=dict(title='Month'),
    yaxis=dict(title='Number of Visitors'),
    margin=dict(l=100, r=100, t=100, b=100)
)

fig = go.Figure(data=steps, layout=layout)
fig.show()

关键说明 #

• 手动计算百分比保证了每个数值都是相对于当月visit_site_month的占比，精准符合需求
• 通过列表推导式生成自定义文本，可灵活调整显示格式（比如修改小数位数、添加单位等）
• textinfo='text'指定Plotly使用我们自定义的文本内容，替代默认的自动生成信息

内容的提问来源于stack exchange，提问作者Joker221