Topic models of India’s corpus.

The present study investigates air pollution dynamics through newspaper discourse in India and Pakistan, where both countries rank among the top five countries affected by air pollution. The study focused on newspaper discourses over almost two decades (2005–2023). The study applied Latent Dirichlet Allocation (LDA), a robust algorithm for analyzing the large text corpus. The study underpinned Computational Grounded Theory, which relies on the fact that computation is a way to reveal the hidden meanings beyond the text. The LDA-generated topics reveal that both countries face the toxicological effects of air pollution on health. The primary topics extracted through LDA revolve around discourses related to vehicular emissions, industrial emissions, and urbanization. In addition, the control measures taken by both countries relate to emission standards. The study also has implications for policymakers and planners considering these directions to control air pollution.