使用VaderSentiment分析Twitter数据集时遇TypeError问题求助

错误原因 #

触发TypeError: 'float' object is not iterable的核心原因是：PreProcessed_Tweets列中存在浮点数类型的值（通常是NaN空值），而Vader的polarity_scores()方法仅支持字符串输入，无法处理浮点数类型的数据。

修复步骤 #

1. 预处理数据，统一列类型 #

先将PreProcessed_Tweets列转换为字符串类型，同时把NaN空值替换为空字符串，避免类型冲突：

# 转换为字符串并处理NaN
df['PreProcessed_Tweets'] = df['PreProcessed_Tweets'].astype(str).replace('nan', '')

2. 优化循环逻辑，减少重复调用 #

原代码每次循环重复调用4次polarity_scores()，既低效又冗余，改为单次调用后提取所有分值：

scores = []
analyzer = SentimentIntensityAnalyzer()

for tweet in df['PreProcessed_Tweets']:
    # 单次调用获取所有情感分值
    sentiment = analyzer.polarity_scores(tweet)
    scores.append({
        "Compound": sentiment["compound"],
        "Positive": sentiment["pos"],
        "Negative": sentiment["neg"],
        "Neutral": sentiment["neu"]
    })

sentiments_score = pd.DataFrame.from_dict(scores)
df = df.join(sentiments_score)
df.head(20)

3. 可选：跳过空内容的推文 #

如果不需要对空字符串进行情感分析，可以在循环中添加判断，为空内容填充NaN：

scores = []
analyzer = SentimentIntensityAnalyzer()

for tweet in df['PreProcessed_Tweets']:
    # 跳过空内容
    if not tweet.strip():
        scores.append({
            "Compound": np.nan,
            "Positive": np.nan,
            "Negative": np.nan,
            "Neutral": np.nan
        })
        continue
    sentiment = analyzer.polarity_scores(tweet)
    scores.append({
        "Compound": sentiment["compound"],
        "Positive": sentiment["pos"],
        "Negative": sentiment["neg"],
        "Neutral": sentiment["neu"]
    })

简洁替代方案：使用pandas apply方法 #

可以用apply替代循环，代码更简洁高效：

def get_sentiment(tweet):
    if not isinstance(tweet, str) or not tweet.strip():
        return pd.Series([np.nan, np.nan, np.nan, np.nan])
    sentiment = analyzer.polarity_scores(tweet)
    return pd.Series([sentiment['compound'], sentiment['pos'], sentiment['neg'], sentiment['neu']])

df[['Compound', 'Positive', 'Negative', 'Neutral']] = df['PreProcessed_Tweets'].apply(get_sentiment)

内容的提问来源于stack exchange，提问作者susne