AidData researchers have been busy in a number of countries testing how remote sensing, machine learning, and social network analysis can both measure and increase growth and consumption of a crop with vital and proven health benefits: orange-fleshed sweetpotatoes (OFSP).
Sweetpotato—despite its name being a root, not a tuber—is a superfood: the beta-carotene that lends it its color also makes it high in vitamin A. Deficiency in this vitamin is associated with disease and death from common childhood infections and is the world's leading cause of preventable childhood blindness. It also contributes to maternal mortality. Lack of vitamin A is a significant problem across parts of sub-Saharan Africa, South Asia, and elsewhere.
There are other advantages to planting OFSP: “Improved varieties of orange-fleshed sweetpotatoes are heat- and drought- tolerant, meaning they can improve food security in areas that are increasingly vulnerable to climate change,” says Katherine Nolan, AidData Research Scientist, who herself has a small sweetpotato patch in her garden in Williamsburg, Virginia, where AidData, a research lab at William & Mary, is based. “This in addition to the tremendous nutritional benefits.”
With organizational partners from across the globe, Nolan and other researchers at AidData are conducting or have recently conducted three evaluations across Africa: a retrospective geospatial impact evaluation (GIE) on the sustainability of an emergency OFSP and potato program in Ethiopia; a survey-based analysis of the impact of gendered social networks on the spread of agricultural information in farming communities in Ghana; and using remote sensing to measure the spread of sweetpotato in communities in Tanzania. (More on these projects below).
AidData’s OSFP program, in partnership with the International Potato Center (CIP), is tracking how information on sweetpotato farming—as well as the roots, seeds and vines themselves—spreads. And can the adoption of this vital crop be observed and measured from the sky or from space? Using remote sensing and Earth Observation tools allows our researchers to not only accurately see the extent of OSFP plantings—but also who is missing out or underserved, whether farmers or potential consumers.
“Lessons from our evaluations are meant to assist farmers and aid organizations plant more sweetpotatoes and get more families incorporating sweetpotatoes into their diets. We ultimately want to help farmers become more climate-resilient through the introduction of hardier new varieties currently being developed and related innovations, such as storing tubers in sand over the dry season,” says Nolan.
New and emerging technologies and approaches can help: AidData plans to use satellite imagery and other data sources to enhance machine learning processes for remotely measuring sweetpotato plantings across multiple countries in Africa. Contingent on funding, AidData will also explore how best to provide frequent, real-time, remotely-derived data to farmers and implementers. Among the projects in development:
- AidData will test how satellite imagery can be used to expand and refine coverage of sweetpotato planting measurements. Combined with additional data sources, our researchers will refine the machine learning process to raise the accuracy of our AI models and make their outputs increasingly useful for remotely measuring sweetpotato planting across multiple countries in Africa.
- AidData will use remote sensing data to provide real-time information to farmers and implementers on crop health and yields, and then measure the impact of up-to-date, frequent agricultural data on the uptake of OFSP programs. This fast feedback could provide additional benefits to farmers who may be less likely to reach out for or obtain assistance, such as marginalized individuals or women.
- AidData will build on an early evidence base that looks at the influence of gendered social networks and the gender of program promoters on the uptake of agriculture interventions and technology, and how these affect both female and male farmers.
While sweetpotatoes are the crop being studied, the boundary-pushing techniques employed by our researchers are replicable for other crops and agricultural interventions too. But orange-fleshed sweetpotatoes are a high priority for many populations, governments, and aid organizations—and rightly so. In addition to vitamin A, they are also a source of Vitamins B, C and E, as well as potassium and manganese. Beyond the roots, the leaves and vines are also edible and packed with nutrients, making the entire plant valuable.
Scientists at CIP, our partner in this work, have also developed breads where 45-50% of the wheat flour has been replaced with OFSP puree. Bakers in developing countries that adopt the puree can lessen their reliance on costly imported flour, reduce the amount of sugar and oil used in their recipes, and still generate all the vitamin A benefits.
CIP, which is headquartered in Lima, Peru, is working to get the word—“sweetpotato”—out. And it is one word, not two, despite what some spellcheckers will tell you.
In Ethiopia, AidData has worked with the Pulte Institute for Global Development at Notre Dame and the non-profit research organization Mathematica to evaluate an emergency humanitarian assistance program CIP implemented in 2021. The program distributed potatoes and OFSP to farmers alongside nutrition and agriculture training, with a specific focus on families with young children. Our researchers found that a significant number of farmers in program areas retained specific agricultural information about OFSP, and almost 20% of them still cultivate the variety, despite the many challenges they’ve faced since the intervention. There was also a positive impact on knowledge and consumption preferences for and greater consumption of OFSP in the treatment areas.
In Ghana, AidData, in partnership with CIP, is in the midst of a project evaluating how best to encourage rural women and men to join “Growing Futures Clubs” (GFCs), where they learn sustainable sweetpotato cultivation practices. CIP and AidData together designed a randomized controlled trial in which communities were randomly assigned either video or traditional informational sessions, as well as “community-based extension agents” (farmers who are trained to support other farmers in agricultural production) who differ by gender and leadership status—all with the goal of assessing how the person and mode of delivering this information influences its uptake. AidData researchers also collected household-level information on spouses’ social networks to pair with baseline survey data and a yet-to-be collected endline survey, in order to measure any differences between sweetpotato uptake and the intensity of a farmer’s social ties.
This recently-completed social network survey interviewed wives and husbands separately to reveal intra-household network discrepancies and overlaps. The survey found significant differences between spouses’ social networks, with approximately 80% of the husband’s social network and 75% of the wife’s social network being distinct from each other. Full results will be published in 2026.
Meanwhile, in Tanzania, CIP recently began the implementation of a large-scale project that includes educating extension agents on sharing information about OFSP, distributing OFSP vines and roots, holding nutrition seminars and cooking demonstrations, and strengthening the business case—both the supply and demand side—for OFSP as a cash crop. AidData is partnering with CIP and the American Institutes for Research (AIR) for a prospective evaluation of the project, similar to the study in Ghana described above. Like in Ghana, the program in Tanzania randomly assigns communities to have different ratios of male and female community-based extension agents, in order to see how the gender and other characteristics of these educators impact the uptake of information about OFSP and the adoption of new agricultural practices.
AidData and CIP researchers are exploring additional research questions, including the feasibility of mapping the social networks of farmers to understand how different types of relationships—such as those that farmers have with close friends versus acquaintances with whom they talk to about farming or nutrition—affect the spread of sweetpotato across a community. Another intriguing angle: analyzing whether the production of OFSP by smallholder farmers spreads more in communities where sweetpotato commercialization is actively being promoted. Either way, it all comes back to one word: sweepotato.
