Research on the Application of Social Media Big Data Analysis in Cultural Communication

Abstract

Against a digital-era backdrop, social media platforms generate large volumes of cultural communication data. Existing scholarship largely concentrates on technical demonstrations yet lacks theory-grounded large-sample empirical analysis. Practical challenges including imprecise audience targeting and inadequate effect evaluation arise from absent multi-dimensional evaluation frameworks. This study conceptualizes cultural communication as a dynamic interaction between communicators and audiences. With the help of Extreme Gradient Boosting-SHapley Additive exPlanations (XGBoost-SHAP), emotional scoring, and social network analysis, this paper studies the multi-platform social media data set of millions. Empirical results identify the gradient-diffusion pattern of cultural hotspots, the quantifiable optimal parameters of short video content, and the stronger cross-circle communication ability of vertical small and medium-sized bloggers. This study constructs a theory-driven analysis process to provide feasible strategies for data-oriented cultural communication.