Scrape and Analyze Maoyan Movie Reviews for “Ne Zha” with Python

Data Acquisition

Open the Maoyan page for the movie, switch the browser to mobile mode, and scroll to the review section. The reviews are loaded dynamically via XHR requests; inspecting the network panel reveals a JSON endpoint that contains all comment data.

Locate the XHR request and copy its URL. The response is a JSON array with fields such as gender, level, city, rating, and comment text.

Use a simple Python script (e.g., requests + BeautifulSoup) to fetch the JSON, iterate over the items, and save the result to a local .json file.

Data Analysis

After collecting roughly 5,000 comments, load the JSON into a pandas DataFrame. Extract useful columns: gender (0 unknown, 1 male, 2 female), user level, city, rating, and the comment text.

Review Visualization

Overall Rating – The movie scores an average of 9.7 on Maoyan, with 86.5% of users giving a perfect score.

Gender Distribution – Among reviewers who disclosed gender, the male‑to‑female ratio is roughly 1:1, with a slight female majority.

City Distribution – A histogram shows the top ten cities by comment count, while a dynamic map visualizes the 80 cities with the most reviewers.

Comment Word Cloud – Using the wordcloud library, the most frequent positive words are “好看”, “不错”, “特效”, while narrative‑related words like “剧情” and “故事” also appear prominently.