The digital transformation of traditional media outlets has become a critical imperative in the evolving media landscape. As traditional media grapple with the challenges of adapting to the rise of digital and social media platforms, understanding the factors that shape their performance and user engagement on these emerging channels is of paramount importance. This study examines the comparative performance of a traditional media account with a government background (Hubei Daily) and a specialized new media account (Cover News) on Douyin platform, with a focus on the distribution of emotional sentiment and thematic diversity in user-generated content. We collected 19,343 comments from “Hubei Daily,” with 18,792 valid (97.15%), and 36,737 comments from “Cover News,” with 28,910 valid (78.69%). Our research used a BERT fine-tuned model for sentiment analysis, TF-IDF for keyword extraction, keyword co-occurrence analysis, and a large language model for topic modeling. This approach allowed us to capture a wide range of discussion topics and understand the thematic focus. Findings show that the new media account covered broader topics, whereas the traditional media account had a more concentrated thematic focus. Typically, such events should generate negative emotions and discussions around accountability. However, the study indicates that users are more cautious in their comments on official media, suggesting ineffective digital transformation by traditional media. This underscores the need for traditional media to better understand and adapt to content preferences and engagement patterns of shorts platform users, highlighting the complexities of transitioning to emerging platforms and the importance of effective content strategies and meaningful user engagement.