PhD candidate at the Wits School of Public Health and co-investigator at Wits Planetary Health Research, University of the Witwatersrand. Based in Cape Town, South Africa.
I work on climate and health in African cities, with a focus on heat exposure and health outcomes in Johannesburg. Most of my time goes to the HE²AT Center, an NIH-funded U54 DS-I Africa initiative running across Johannesburg and Abidjan, where I work on data science and research data governance.
Heat vulnerability and urban thermal exposure in African cities, linking cohort and clinical data with satellite and reanalysis climate products. Current work covers distributed lag non-linear models for temperature and health outcomes, explainable machine learning (SHAP, gradient boosting) for biomarker screening, geographically weighted PCA for ward-level vulnerability, and Local Climate Zone classification.
Python and R for epidemiological modelling: DLNMs, case-crossover and time-series designs, survival analysis. Google Earth Engine and the usual geospatial stack for remote sensing and spatial analysis. Azure and Databricks for harmonising multi-site health data and running governed cloud research infrastructure to NIH and institutional requirements.
- dlnmpy: distributed lag non-linear models in Python, validated against the R
dlnmpackage - HE2AT-RP2-Metadata-Catalogue: metadata catalogue for the multi-system biomarker heat analysis in Johannesburg
- LCZ_coding: Local Climate Zone and temperature extraction workflows for Johannesburg (Earth Engine, ERA5, MODIS LST)
- admire-cohort-map: study site mapping for the ADMIRE cohort
- ORCID: 0000-0003-2517-929X
- Joburg HeatWatch, a heat-health early warning prototype: joburg-heatwatch.web.app
- Email: craigparker6@gmail.com
Open to collaboration on climate-health analytics, heat early warning systems, and reproducible research data management.