USING ECO-LDA AS A FRAMEWORK FOR THE COMPUTATIONAL ECOLINGUISTIC ANALYSIS OF PAKISTANI MEDIA COVERAGE OF THE LAHORE SMOG (2024-25): A RECENT SNAPSHOT OF THE CRISIS

Abstract

Climate change is a long-lasting environmental issue in Southeast Asian countries, specifically Pakistan. 
This study investigates how Pakistani media has framed the Lahore Smog crisis in recent years (2024-25). 
In addition, the study aims to understand the ecological relationship framed by the Pakistani media, and 
how has the framing itself evolved over the last two years. Two major Pakistani News outlets, including 
Dawn and The News International, were selected for data collection. 200 articles were purposively 
collected for the corpus 2024, while 160 articles accounted for the 2025 corpus. Moreover, the study 
proposed a novel and replicable Eco-LDA framework for data analysis, which blended Latent Dirichlet 
Allocation (LDA), Frame Semantics, and Stibbe’s (2015) Ecolinguistic framework. The Eco-LDA 
framework was specifically designed to identify latent frames in the corpora and assess the ecological 
visibility in media discourse.  
Results reflect a shift in media framing in the recent years. In 2024, media framed the Lahore smog as 
anthropocentric and episodic issue; it highlighted immediate health risks and civic disruptions while 
backgrounding the root-causes of the ecological crisis. In 2025 media framing, on the other hand, industrial 
emissions, crops burning, climate change, and policy interventions were foregrounded. The findings reveal 
an increasing ecological literacy in the news media discourse in Pakistan over 2024 and 2025.  
The research provides latest insights into the evolving media framing of the Lahore smog. It may assist 
environmental communication, climate polices, governance, and public ecological understanding. 
Moreover, methodologically, this research contributes to computational ecolinguistic inquiry by 
introducing a replicable Eco-LDA framework.