A Machine Learning Analysis of Climate Change & Human Health Projections in Pakistan
Author: Zarak Jamal Khan


This thesis explores the profound challenge of global warming and climate change in Pakistan, focusing on deterioration of health. Through the utilization of machine learning techniques on climate change and causes of death datasets and is set to investigate the linkages among drivers of the climate change specifically in the context of Pakistan. The analysis reveals robust correlations between climate change and victims of respiratory diseases, while the associations with the victims of digestive problems and cardiovascular diseases are found to be comparatively less significant.

The examination of causality emerges as a potential solution to overcome the limitations of current machine-learning approaches. The interdisciplinary nature of causality, drawing from fields such as epidemiology, economics, statistics, and computer science, underscores the significance of collaboration and knowledge exchange. The research focuses on one of the fundamental task that is causal discovery. By employing causal discovery tools, the study delves into investigation and exploration of the causal linkages between climate change and human deaths, identifying both direct and indirect relationships with the drivers of climate change and leading causes of mortality in Pakistan.

While the study provides valuable insights into the intricate relationship between climate change and human health, further comprehensive analysis and extensive data are needed to obtain more precise and accurate results. The thesis emphasizes the necessity of a multidisciplinary approach to deepen our understanding of causality and climate change’s health implications, leading to evidence-based policies and interventions.

In summary, this thesis underscores the urgent need to address climate change as a critical issue in Pakistan. By unraveling the correlations between climate change and the victims of respiratory diseases, it contributes to the existing body of knowledge. The research highlights the importance of causality in comprehending complex phenomena, advocates for cautious interpretation of correlations, and demonstrates the potential of causality in addressing the limitations of machine learning. By further exploring the causal pathways and gathering extensive data, the thesis aims to enhance our understanding of the relationship between climate change and human health, paving the way for effective strategies to safeguard the wellbeing of the population in Pakistan.

Meta Data

Keywords : & NASA-GISS-E2-1H, Climate Change, CMIP6, Forecasting, Machine learning, Model
Supervisor: Amena Urooj

Related Thesis​