PreEnvBD Preterm Birth and Environment Dataset

The PreBEnvBD dataset is a novel dataset designed to support the analysis and prediction of preterm birth in Bangladesh. It integrates maternal and child health data with environmental and meteorological factors to enable comprehensive, multidisciplinary research. The dataset has been constructed using maternal and child health information derived from three components of the Bangladesh Demographic and Health Survey (BDHS): Household, Fertility and Maternal & Child Health datasets. These data were obtained through direct correspondence with survey administrators and carefully processed for research use. To enhance analytical depth, the health data are systematically mapped with environmental and climate related variables. Meteorological parameters including rainfall, humidity and temperature are collected from the Bangladesh Meteorological Department (BMD). In addition, air pollution indicators such as PM₂.₅, PM₁₀, CO, SO₂, NO₂, and O₃ are measured using an IoT based monitoring system deployed across six high density residential areas of Dhaka: Dhanmondi, Kuril Bishwa Road, Mirpur, Mohakhali, Jatrabari and Uttara. This ensures diverse environmental exposure representation across urban settings. The final dataset contains approximately 73,000 records and 58 attributes. Each record represents a comprehensive combination of maternal, neonatal, household and environmental characteristics. Key maternal and neonatal variables include gestational duration, maternal age, education level, anthropometric measures (height, weight, BMI), delivery type, neonatal survival status and birth order. Household level features include wealth index, drinking water source, sanitation facilities, cooking fuel type and housing density. Environmental attributes capture both air quality indicators and meteorological conditions.