Enhancing Semantic Understanding by Bridging Topic Modeling and Thematic Analysis: An Empirical Study on Self-Help Twitter Corpus and In-Depth Interviews

Bridging data science with social science is useful for social scientists to develop a more transparent approach to their data analysis and enhance the rigor and credibility of qualitative research. This study aims to offer the potential of an iterative process which allows to discover some patterns through exploratory analysis of big Twitter data drawn from four influential self-help gurus’ accounts and grounding these patterns into smaller but richer interview data based on 24 in-depth interviews conducted among self-help consumers. Going in between the findings from topic modeling and thematic analysis helped triangulate the analysis of the same research question from multiple data sources and from different methodological perspectives.