What Do Netizens Really Think of Bilibili’s “Post‑Wave” Video? Sentiment & Word‑Cloud Analysis

Data Collection

Comments were scraped from three sources: a Zhihu answer discussing the 2020 Youth Day video “后浪”, the Weibo post under the hashtag #献给年轻一代的演讲#, and Bilibili danmaku for the same video.

Sentiment Analysis

The Baidu AI NLP service was used to classify the sentiment of Weibo and Bilibili comments. The Zhihu answer was too long for reliable sentiment classification, so it was omitted.

from aip import AipNlp
""" 你的 APPID AK SK """
APP_ID = '你的APP_ID'
API_KEY = '你的API_KEY'
SECRET_KEY = '你的SECRET_KEY '
client = AipNlp(APP_ID, API_KEY, SECRET_KEY)
text = "XXXXXXXX"
""" 调用词法分析 """
response = client.sentimentClassify(text)
# "sentiment":2,    // 0: negative, 1: neutral, 2: positive
# "confidence":0.40, // confidence score
# "positive_prob":0.73, // probability of positive
# "negative_prob":0.27 // probability of negative
for info in response['items']:
    if info['sentiment'] == 2:
        print("正向")
    if info['sentiment'] == 0:
        print("负向")
    if info['sentiment'] == 1:
        print("中性")
    print("可信度：", info['confidence'])
    print("属于积极类别的概率是：", info['positive_prob'])
    print("属于消极类别的概率是：", info['negative_prob'])

Note: the phrase “奔涌吧，后浪！” is interpreted as a negative sentence by Baidu AI and therefore requires special handling.

Results show that Bilibili danmaku are predominantly positive, while Weibo comments are mainly negative.

Word Cloud Generation

High‑frequency terms from each platform were extracted and visualized as word clouds.

from scipy.misc import imread
from wordcloud import WordCloud, ImageColorGenerator

isCN = 1  # 0: English segmentation, 1: Chinese segmentation
back_coloring_path = '浪花.jpg'  # background image path
text_path = 'reviews.txt'        # text file path
stopwords_path = 'stop_word.txt' # stop‑words list
imgname1 = '词云图.png'          # output image name
back_coloring = imread(back_coloring_path)

wc = WordCloud(
    font_path='C:\Windows\Fonts\simfang.ttf',
    background_color='white',
    max_words=3000,
    mask=back_coloring,
    max_font_size=200,
    min_font_size=20,
    random_state=42,
    width=2000, height=1720, margin=4)
words = {}
for i in word_counts:
    words['{}'.format(i[0])] = i[1]
wc.generate_from_frequencies(words)
plt.figure()

Observations

Weibo and Zhihu users focus on topics such as “reality”, “society”, “housing prices”, and “韭菜”, expressing cynicism toward the motivational video. In contrast, Bilibili users celebrate the platform’s growing influence and view the video positively.