Ongoing work
Journal article · Main author Submitted to NHESS · Oct 2026
Spatial transferability of knowledge-based, statistical and machine learning models for flood susceptibility mapping in Bangladesh
With Khandakar Hasan Mahmud. Building on my BSc dissertation, this study compares an expert-based model (AHP), a frequency ratio model and three machine learning models for mapping flood susceptibility across Bangladesh, using ten conditioning factors and flood records from the Global Flood Database. It tests how well each model predicts in regions it was not trained on, and finds that the simpler frequency ratio model transfers better than machine learning while staying transparent.
Research in progress In progress
Integrating flood hazard and road-network exposure to map flood risk to urban roads: a Fuzzy AHP approach in Dhaka, Bangladesh
Combines a flood hazard model with the exposure of Dhaka's road network to show which city roads face the greatest flood risk, using Fuzzy AHP to weight the factors.
Journal article · Sole author In preparation
Socio-economic conditions and perceived microclimatic stress among tea garden workers: a case study of Khadimnagar Tea Estate, Sylhet
Based on surveys with tea garden workers during my master's fieldwork in Sylhet, this study looks at wages, housing and support, and at how workers experience heat and humidity in the gardens. It finds an overlapping burden of low pay, poor housing and heat exposure at work, and points to wage reform, better housing, shade trees and heat-safe work hours.
