Reading Between The Reels: A Data Analysis of City Entertainment Magazine
This project, authored by CHAN Yin Kei and YANG Run Xin, examines how one publication serves as a repository for social and industrial change. We focus on City Entertainment Magazine (1979–2007), a popular Hong Kong periodical. This magazine documented film, television, music, and popular culture for nearly thirty years. Our study analyses a corpus of 25,266 articles. We use computational tools to reveal large-scale patterns. By combining natural language processing with humanities frameworks, we trace shifts in Hong Kong’s identity. This work shows how digital methods enhance traditional cultural analysis.
The Archive as Cultural Memory
City Entertainment Magazine can be understood as more than a specialised publication focused on film; it functions as a longitudinal cultural archive that captures the evolution of Hong Kong’s media ecology. Published bi-weekly between 1979 and 2007, the magazine spans a period widely recognised as the rise and transformation of Hong Kong’s entertainment industries, particularly its globally influential cinema.
Across its issues, the magazine engages not only with film production and criticism, but also with television broadcasting, the recording industry, celebrity culture, and broader forms of social commentary. As such, it provides a multifaceted account of cultural production and reception, offering insights into both industrial structures and everyday cultural discourse.
Dataset Significance
The analytical potential of this project derives from the unusual completeness and structure of the dataset. The corpus consists of 725 fully digitised issues comprising 25,266 articles written by 3,640 distinct contributors over a continuous 28-year period. Each article is accompanied by structured metadata, enabling systematic analysis across time.
Such continuity is rare in cultural archives, which are often fragmented or incomplete. In contrast, this dataset allows for longitudinal investigation at scale, making it particularly well suited to computational approaches that seek to identify patterns, trends, and transformations across extended temporal spans.