From Words to Winners: Investigating Liu Cixin’s Bestseller Textual Features with Computational Methods
2025 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE credits
Student thesis
Abstract [en]
This study investigates the textual factors behind the global success of Liu Cixin’s The Three-Body Problem trilogy by comparing its themes, plot, and style with his non-bestselling works. Using digital humanities’ distant reading approach, it analyzes all English translations of Liu’s novels (approximately 1.24 million words) with digital tools: Latent Dirichlet Allocation (LDA) for topic modeling, TextBlob for sentiment analysis, and spaCy for stylometric analysis.
Findings show minimal textual differences. The trilogy’s themes lean slightly toward hard science fiction, its plot aligns with typical bestseller patterns, and its style is highly consistent with non-bestsellers. Its exceptional market success is largely attributed to contextual factors, including Tor Books’ strategic publishing, the 2015 Hugo Award, cross-cultural narratives, reader engagement, and film/TV adaptations, amplified by a positive feedback loop.
This study confirms the potential and limitations of digital humanities in single-author bestseller research and, using the trilogy as a case study, discusses how textual quality and market mechanisms together drive bestseller success.
Place, publisher, year, edition, pages
2025. , p. 79
Series
Theses within Digital Humanities ; 53
Keywords [en]
Computational text analysis, topic modeling, sentiment analysis, stylometry, bestseller, Liu Cixin, The Three-Body Problem
National Category
Interdisciplinary Studies in Humanities and Arts Natural Language Processing General Literature Studies Media and Communication Studies
Identifiers
URN: urn:nbn:se:uu:diva-562544OAI: oai:DiVA.org:uu-562544DiVA, id: diva2:1979057
Subject / course
Digital humanities
Educational program
Master's Programme in Digital Humanities
Supervisors
Examiners
2025-07-012025-06-302025-07-01Bibliographically approved