Public sentiment module of project SmartCity · Used the Python lib (BeautifulSoup4) to scrape reviews and scores of 11 scenic spots on dominant travel sites with concurrent.futures.ThreadPoolExecutor to accelerate · Added a module(scenic name->opinions on every spots & sites)by scraping params of spots and its names (Selenium) · Optimized code, increased fault tolerance, implemented dynamic-update and interface calling with Flask · Cleaned and structured over 100k data into MySQL by PyMySQL lib · Utilized THULAC(Python) for Chinese text segmentation, NLP component developed by AI team for sentiment classification to analyze the popularity, and compared the results with original scores of travel sites · Participated in using React to visualize the data such as word frequency cloud and sentiment classification