如何用Python和Pandas分组统计DataFrame并提取Top3关键词？

完全可以用Python + Pandas实现这个需求，核心是通过分组聚合完成统计和文本词频分析，具体步骤如下：

1. 导入依赖库并准备数据 #

import pandas as pd
from collections import Counter
import string
from nltk.corpus import stopwords
import nltk

# 下载英文停用词（首次运行需执行）
nltk.download('stopwords')
stop_words = set(stopwords.words('english'))

# 构造示例DataFrame
data = {
    'ID': [1,2,3,4,5],
    'name_x': ['xx','xy','xz','xu','xi'],
    'st': ['us','us1','us','us2','us1'],
    'string': [
        'Being unacquainted with the chief raccoon was harming his prospects for promotion',
        'The overpass went under the highway and into a secret world',
        'He was 100% into fasting with her until he understood that meant he couldn\'t eat',
        'Random words in front of other random words create a random sentence',
        'All you need to do is pick up the pen and begin'
    ]
}
df = pd.DataFrame(data)

2. 定义文本处理与词频统计函数 #

这个函数负责清洗文本、过滤无意义词汇，并返回每组文本中频率最高的前3个关键词（不足3个时补空字符串）：

def get_top3_words(text_series):
    # 合并组内所有文本内容
    combined_text = ' '.join(text_series.astype(str))
    # 去除标点并转小写，统一词汇格式
    translator = str.maketrans('', '', string.punctuation)
    clean_text = combined_text.lower().translate(translator)
    # 分词
    words = clean_text.split()
    # 过滤停用词（如the、was）和纯数字
    filtered_words = [word for word in words if word not in stop_words and not word.isdigit()]
    # 统计词频
    word_counts = Counter(filtered_words)
    # 提取前3个高频词，不足3个补空
    top_words = [word for word, _ in word_counts.most_common(3)]
    while len(top_words) < 3:
        top_words.append('')
    return top_words

3. 分组聚合生成结果 #

通过groupby按st分组，同时完成name_x计数和关键词提取：

# 分组处理每个组
result = df.groupby('st').apply(
    lambda group: pd.Series({
        'name_x_count': group['name_x'].count(),
        'top1_word': get_top3_words(group['string'])[0],
        'top2_word': get_top3_words(group['string'])[1],
        'top3_word': get_top3_words(group['string'])[2]
    })
).reset_index()

# 查看最终结果
print(result)

结果示例 #

运行后会得到符合要求的DataFrame：

注：如果想更精准处理缩写（比如couldn't），可以将分词逻辑替换为nltk.word_tokenize，基础split已能满足大部分场景需求。

内容的提问来源于stack exchange，提问作者dd99