I School Ph.D. student Susana Constenla-Villoslada, Professor Joshua Blumenstock, and collaborators at Kenya’s National Drought Management Authority (NDMA), Cornell University, and UC Santa Barbara, have been awarded a grant by the Jameel Observatory to implement a machine learning-based early warning system for malnutrition crises among children under five in Kenya.
Over eighty percent of Kenya is made up of arid and semi-arid lands (ASAL). These lands are especially sensitive to the effects of climate change which have manifested through severe recurring droughts over the past two decades. In 2011, the Kenyan government created the NDMA to monitor and address the impacts of these long-standing droughts and collect population data for relief efforts. Part of this data collection is gathering middle upper arm circumference (MUAC) measurements of children under five years old to monitor children’s nutritional status. However, despite collecting this data for years, it has not yet been utilized for any needs assessment of humanitarian aid related to climate-related hunger crises.
Constenla-Villoslada and her collaborators at Cornell University and UC Santa Barbara sought to change that. Constenla-Villoslada started working on the project while working as a research analyst for the International Food Policy Research Institute in Washington D.C.
“It seems to me that [hunger prevalence in Africa South of the Sahara] doesn’t get enough attention anymore,” she said. “The consequences for the kids are huge. There are developmental consequences of being malnourished as a kid…It’s unacceptable that as of 2023, around ten million kids in the Horn of Africa were malnourished.”
Using unprocessed MUAC data as a starting point, she started building a machine learning-based model to predict malnutrition prevalence among these children during her time in D.C. After three years of development, this predictive model will be prototyped in Spring 2025 as a new addition to NDMA’s drought monitoring efforts. The results of the model will be incorporated into a dashboard that will allow the NDMA and other stakeholders to monitor the level of child malnutrition in ASAL Kenyan communities and ideally alert impending hunger crises. That alert will hopefully trigger an intervention where the NDMA will provide aid through existing social safety net programs to those in need of food and other supplies.
After working in tandem with the NDMA for the last three years, the grant by the Jameel Observatory will allow Constenla-Villoslada and Professor Blumenstock to materialize the research on the ground, with upcoming trips to Kenya next Spring where they will finalize deployment details by incorporating NDMA’s feedback and carry out technology transfer to the organization. The Jameel Observatory, which is funding research and travel expenses for this purpose, is an organization based at the University of Edinburgh focused on research related to the impacts of climate change in low and middle-income countries, with a focus on vulnerable communities in the Horn of Africa.
“With its multidisciplinary approach, the I School is helping create the space where a new generation of researchers working in international development can develop the necessary skills to provide new solutions to ongoing issues.”
“This grant is impact-oriented. You need to be in partnership with someone on the ground; they want innovations on how to translate early warning predictions into actions…In our case, I think our project got attention because it’s currently very rare to have an early warning system that predicts an actual targetable measure [malnutrition prevalence among children under five],” Constenla-Villoslada explained. “[The Jameel Observatory] see us as potential innovators, as people that are actually carrying things that can be applicable and translate research into action.”
“Our collaborators in Kenya are great professionals with a lot of expertise who have helped us so much during this whole process. Part of our collaboration with them entails carrying out technology transfer of the model, so that they can run it on their own moving forward. Capacity building and technology transfer with research partners in developing economies is, in my view, something universities should do as much as possible,” Constenla-Villoslada said. “With its multidisciplinary approach, the I School is helping create the space where a new generation of researchers working in international development can develop the necessary skills to provide new solutions to ongoing issues.”
Constenla-Villoslada gives special thanks to a crucial project team member, Professor Chris Barrett from the Dyson School of Applied Economics and Management at Cornell University’s SC Johnson College of Business, for starting this initiative and for his ongoing support and guidance; the whole research team thanks their partners at the National Drought Management Authority, Clinton Ouma and Nelson Mutanda, for giving the team access to the data needed for this project, for providing feedback throughout the whole process, and for their continued support and excitement on making this research project become a reality.
