Forecasting and Quantifying Event-Induced Stress during the Prayagraj Maha Kumbh Using Machine Learning and a Composite Event Stress Index

Large-scale public events create short-term but intense environmental and social pressures that challenge urban planning and sustainability goals. These pressures and future expectations are critical to sustainable event planning and mitigation of risks. This study is a time-series forecasting, machine-learning analysis, and composite index construction built into a unified framework exploring four Maha Kumbh cycles that have taken place in Prayagraj (1989, 2001, 2013, and 2025). Forecasting models are used to predict future trends in air pollution, water quality, waste generation, crime incidents, visitors, and sentiments. The relative impact of environmental and social drivers is determined using machine-learning methods, in particular, Random Forest, and transformed into objective weights to create an Environmental-Social Stress Index (ESI). Normalization and aggregation of all indicators produce event-year comparative stress scores. The proposed framework provides a logical method of assessing environmental and social stress in massive gatherings. The analysis of forecasting, data-driven weighting, and composite indexing can inform policymakers in taking action to enhance preparedness, resource allocation, and sustainability plans of future Kumbh cycles and other mass gatherings.