Towards nuanced and critical data analysis: an exploratory investigation into complexities of Goodreads reviews for children's books

Abstract

Introduction. With the advance of artificial intelligence and big data, digital humanities have been empowered with new research affordances. However, computational modelling on cultural datasets without a critical examination of the datasets risks decontextualized datafication and misinterpretation, which does not necessarily advance our knowledge into humanities and social sciences questions. This paper presents an exploratory investigation into 37,408 Goodreads reviews on children’s books to showcase the complexities of cultural datasets online, to inform contextualized and critical cultural data analysis.

Method. We conducted topic modelling and qualitative analysis of content on the reviews retrieved by regular expressions and visualized our findings.

Analysis. Our analysis empirically illuminates (1) participation in reviewing children’s literature online, and (2) presentation of topics and opinions in the reviews.

Results. We found that although various family members across generations participate in reading children’s books and reviewing them online, female members play particularly active roles, and children’s opinions are underrepresented.

Conclusions. In addition to gaining insights into reviews on children’s books, this study also demonstrates the importance of exploratory data analysis in gaining a more nuanced understanding of cultural datasets online, which is essential for advancing insightful and critical data analysis.