Md. Mominul Islam
Economics Researcher
Climate · Development · Inequality
Prospective PhD Researcher
Department of Economics
Jagannath University
Dhaka 1100, Bangladesh
Data & Research Methods
I work with national household surveys, longitudinal datasets and geospatial observations. The information below describes research data sources and analytical approaches; it does not provide restricted microdata or imply that each dataset has been used in every project.
Research Data
Bangladesh Multiple Indicator Cluster Surveys (MICS)
Coverage: 2006, 2012–13, 2019, 2025
Research use: Early home-learning, caregiver participation, inadequate supervision, child development, water microbiology and women’s indicators
Bangladesh Demographic and Health Surveys (BDHS)
Coverage: 1993–94 through 2022 (nine rounds in current women’s life-course work)
Research use: Education, marriage, fertility, maternal health and women’s status
Household Income and Expenditure Survey (HIES)
Coverage: 2022
Research use: Household shocks, consumption, costly coping, food insecurity, labour and social protection
Bangladesh Integrated Household Survey (BIHS)
Coverage: 2011–12, 2015, 2018–19
Research use: Panel household structure, schooling, agricultural production and rural welfare
CHIRPS rainfall
Coverage: 1981–2025, daily and monthly district extracts
Research use: Historical precipitation, extremes and seasonal anomalies
ERA5-HEAT / Universal Thermal Climate Index (UTCI)
Coverage: 2006–2025 local daily/monthly holdings; assessment-time linkage in 2025
Research use: Physiological heat-stress exposure across time and geography
ERA5-Land
Coverage: Historical weather and hydrological series; coverage varies
Research use: Air temperature, soil moisture, runoff and antecedent conditions
VIIRS nighttime lights
Coverage: 2015–2025 district-month panel
Research use: Satellite-observed nighttime radiance as a proxy for local economic activity
MODIS thermal and vegetation products
Coverage: District-month and other extracts; dates vary
Research use: Land-surface temperature, NDVI and EVI
Historical flood and cyclone exposure
Coverage: 1988 flood; Cyclone Sidr in 2007
Research use: Historical hazard classifications and exposure-cohort research
Bangladesh district boundaries and geospatial layers
Coverage: 64-district contemporary geometries; harmonized historical variants
Research use: Spatial joins, climate mapping and regional comparisons
Global Mangrove Watch / GTSM coastal data
Coverage: Mangrove products including 1996–2020; other coverage varies
Research use: Mangrove extent, coastal change, tidal and storm-surge context
Agricultural statistics and rice production
Coverage: Approximately 2015–2024
Research use: District and cropping-season rice productivity, Aus/Aman/Boro
Other satellite indicators
Coverage: Availability and periods vary
Research use: SMAP soil moisture, Sentinel-5P NO₂, land cover and remote sensing
Methods and Tools
My empirical work draws on survey-weighted analysis, applied econometrics, panel-data methods, causal-identification assessment, spatial analysis and reproducible computation. Techniques include:
- Descriptive analysis, sampling weights, stratification, primary sampling units, regression modelling and predictive margins
- Linear, logistic and count-data regression; model diagnostics, confidence intervals and sensitivity analyses
- Fixed-effects panel models, repeated cross-sections, birth-cohort comparisons, distributed lags and event-time specifications
- Difference-in-Differences (DiD) designs and event studies where timing, comparator groups and identifying assumptions can be justified
- Spatial joins, GIS mapping, remote-sensing processing and climate–survey exposure linkage
- Multiple-testing adjustment, missing-data checks, alternative specifications, placebo tests and subgroup analysis
- Structured literature review, evidence mapping, reference verification and reproducible research workflows
Research software includes Stata, SPSS, Python, R, GIS tools and Google Earth Engine. Econometric designs are assessed against their identifying assumptions; association is not presented as a causal effect solely because a coefficient is statistically significant.