Data and Code for "Deep Learning-Based Prediction of Social Media Post Popularity Using Multimodal Content Features"

This repository contains the raw data and analysis code used in the manuscript “Deep Learning-Based Prediction of Social Media Post Popularity Using Multimodal Content Features.” The dataset includes Sina Weibo brand post data used for the empirical analysis, including post text, image information, posting time, and post-level engagement metrics. The code includes scripts for data preprocessing, regression analysis, multimodal deep learning models, baseline model comparison, and robustness checks. The supplementary file S1 Appendix includes sensitivity analyses for popularity-tier definitions, heteroskedasticity and multicollinearity diagnostics, alternative count-model specifications, and quantile regression results.