Plurality Analysis in the Freedom Era Weekly Based on Topic Breadth and Keyword Complexity

This work is based on a historical dataset from before Taiwan's lifting of martial law. Between 1984 and 1989, under the direction of Cheng Nan-jung, The Freedom Era Weekly and the subsequent series magazines were published in over 300 issues under 27 different names. Despite facing severe censorship and bans, they continued to provide extensive current affairs commentary and diverse perspectives on a weekly basis. The Cheng Nan-jung Foundation has digitized the complete texts of these magazines and related political archives, providing valuable historical materials for researching the political spectrum, social dynamics, and historical context before and after the lifting of martial law. This research employs natural language processing techniques to analyze the magazine texts, conducting named entity recognition, keyword extraction, and social network relationship construction. Using diversity quantification methods, it presents the spectrum distribution of topics and issues. Through visualization methods such as association networks, the study examines Taiwan's political discourse before the lifting of martial law from both macro and micro perspectives, analyzing its intensity and changes. This provides objective evidence for understanding the historical roots of contemporary political identity.