求助：如何用Pandas按多列活动数据划分用户活跃度类别

核心思路 #

先计算4个活动水平列的全局均值，再通过逐行判断，将用户分到对应类别：仅某一列高于均值且其余三列均低于均值。

步骤与代码实现 #

假设你的DataFrame列名为sedentary_minutes（久坐分钟）、lightly_active_minutes（轻度活跃分钟）、fairly_active_minutes（中度活跃分钟）、very_active_minutes（高度活跃分钟），按以下步骤操作：

1. 导入依赖库

import pandas as pd
import numpy as np

2. 计算各活动列的均值
   先提取目标列，计算每列的全局均值：

# 提取4个活动水平列
activity_cols = ['sedentary_minutes', 'lightly_active_minutes', 'fairly_active_minutes', 'very_active_minutes']
# 计算每列均值
activity_means = df[activity_cols].mean()

3. 定义分组条件与标签
   用布尔条件匹配每个类别的规则，再对应到分类标签：

# 定义各分类的布尔条件
conditions = [
    # Sedentary：仅久坐数据高于均值，其余均低于
    (df['sedentary_minutes'] > activity_means['sedentary_minutes']) &
    (df['lightly_active_minutes'] < activity_means['lightly_active_minutes']) &
    (df['fairly_active_minutes'] < activity_means['fairly_active_minutes']) &
    (df['very_active_minutes'] < activity_means['very_active_minutes']),
    # Lightly Active：仅轻度活跃数据高于均值，其余均低于
    (df['lightly_active_minutes'] > activity_means['lightly_active_minutes']) &
    (df['sedentary_minutes'] < activity_means['sedentary_minutes']) &
    (df['fairly_active_minutes'] < activity_means['fairly_active_minutes']) &
    (df['very_active_minutes'] < activity_means['very_active_minutes']),
    # Fairly Active：仅中度活跃数据高于均值，其余均低于
    (df['fairly_active_minutes'] > activity_means['fairly_active_minutes']) &
    (df['sedentary_minutes'] < activity_means['sedentary_minutes']) &
    (df['lightly_active_minutes'] < activity_means['lightly_active_minutes']) &
    (df['very_active_minutes'] < activity_means['very_active_minutes']),
    # Very Active：仅高度活跃数据高于均值，其余均低于
    (df['very_active_minutes'] > activity_means['very_active_minutes']) &
    (df['sedentary_minutes'] < activity_means['sedentary_minutes']) &
    (df['lightly_active_minutes'] < activity_means['lightly_active_minutes']) &
    (df['fairly_active_minutes'] < activity_means['fairly_active_minutes'])
]

# 对应分类标签
labels = ['Sedentary', 'Lightly Active', 'Fairly Active', 'Very Active']

4. 添加分类列到DataFrame
   用np.select将条件匹配结果映射到标签，default='Other'处理不符合任何单一条件的用户（比如多列均高于均值的情况）：

df['activity_category'] = np.select(conditions, labels, default='Other')

验证结果 #

可以查看各分类的用户数量，确认分组是否符合预期：

print(df['activity_category'].value_counts())

内容的提问来源于stack exchange，提问作者Teresa