Zimbabwe spans dramatic environmental gradients: the cool Eastern Highlands of Manicaland, the hot dry lowveld of Masvingo and Matabeleland South, and the urban centres of Harare and Bulawayo. These gradients translate directly into differences in heat stress, drought, vegetation cover, and ultimately birth outcomes — patterns that spatial data science can reveal and quantify.
This course is built around the Zimbabwe DHS 2015 birth records — a nationally representative dataset of births collected from women aged 15–49. Working through real geocoded DHS clusters, you will learn how researchers link individual birth outcomes (birthweight, preterm delivery, low birthweight) to satellite-derived environmental exposures at the time and place of pregnancy.
You will learn to construct trimester-specific exposure windows, perform spatial joins between DHS clusters and gridded climate products, and build a logistic regression model predicting low birthweight from heat, drought, and vegetation — all inside pre-configured Jupyter notebooks with no local setup required.
The course closes with a GeoAI assistant that lets you query the full exposure dataset in plain English. You leave with six complete notebooks, reusable geospatial code, and a concrete understanding of how environmental epidemiology is done with real survey data.