Topic ID No.: 2026-G17
Title of research topic:
Transdisciplinary Data Science to Optimise Food System Outcomes and Enhance Resilience under Climatic and Socioeconomic Stress
Description:
UKUDLA is an African–German research centre on sustainable and resilient food systems that conducts transdiscilinary, data-driven research across multiple sites in Southern Africa, combining field experiments, observational studies and stakeholder-driven living laboratories. The project develops data-driven methods to integrate heterogeneous agricultural, environmental, socio-economic and value chain data across the UKUDLA research sites into a unified and comparable representation of food systems. The integrated representation enables cross-site analysis. The research identifies key variables and interactions that determine differences in outcomes, in particular food security, livelihoods and sustainability. It formulates optimisation approaches based on these structures. It extends these approaches to explicitly account for resilience. The goal is robust system behaviour under stress conditions such as climate variability, market shocks and resource constraints. Methodologically, the project combines statistical learning with causal and probabilistic modelling and system-level simulation, including digital twins and data-driven optimisation.
Target region or country (if applicable):
-
Topic background information / scientific relevance:
UKUDLA is an African–German research centre on sustainable and resilient food systems, conducting transdisciplinary, data-driven research across multiple sites in Southern Africa. The sites represent diverse agro-ecological and socio-economic conditions and generate heterogeneous data from field experiments, observational studies and stakeholder-driven living laboratories.
Food systems in Sub-Saharan Africa face increasing pressure from climate variability, market volatility, demographic change and structural inequalities. These pressures affect food security, livelihoods and environmental sustainability. At the same time, large amounts of data are generated across agricultural production, value chains, environmental monitoring and socio-economic surveys. These data remain fragmented, methodologically inconsistent and rarely
integrated across sites or disciplines.
This project addresses this gap by developing data-driven methods to integrate heterogeneous data into a unified and comparable representation of food systems. This enables cross-site analysis and identification of key variables and interactions that determine system performance. Building on this, the project investigates how food system outcomes can be optimised across sites and how differences in performance can be explained.
A central scientific contribution is the explicit incorporation of resilience into optimisation frameworks. The project develops approaches that account for stress conditions such as climate variability, market shocks and resource constraints, enabling robust system behaviour under uncertainty.
The work contributes to data science, systems modelling and food systems research. It addresses a critical methodological gap in the integration and analysis of complex, multi-actor systems and supports evidence-based decision-making for sustainable and resilient food systems.
Research objectives:
Develop methods to integrate heterogeneous agricultural, environmental, socio-economic and value chain data from UKUDLA sites into a unified and comparable food system representation. Identify key variables and interactions that determine differences in food system outcomes across sites, in particular food security, livelihoods and sustainability. Formulate data-driven optimisation approaches based on these structures. Extend these approaches to explicitly incorporate resilience under stress conditions such as climate variability, market shocks and resource constraints. Implement and evaluate the methods using cross-site data and real-world use cases.
Required skills and qualifications of the applicant:
- Strong background in applied data science and quantitative modelling.
- Experience in data engineering and integration of heterogeneous data sources.
- Solid foundation in statistical learning and machine learning, including linear and non-linear models. Experience with causal or probabilistic modelling, simulation methods (e.g. Monte Carlo), and robustness analysis.
- Familiarity with geospatial data and modelling is advantageous.
- Proficient in Python and/or R; experience with scripting and reproducible workflows expected.
- Ability to analyse complex systems and translate results into interpretable insights.
- Experience in scientific visualisation, including interactive approaches, is desirable.
- Excellent English skills and ability to work in interdisciplinary, international research teams.
Contact person and institute in charge:
Prof. Dr. Asis Hallab