Our Research Improving smallholder farmers’ welfare with AI-driven technologies
Field Visits in Karnataka and Bihar. Image Credit: Somya Singhvi & Yanchong Zheng
Field Visits in Karnataka and Bihar. Image Credit: Somya Singhvi & Yanchong Zheng
The research method of this project visualized. Image Credit: Somya Singhvi & Yanchong Zheng
Field Visits in Karnataka and Bihar. Image Credit: Somya Singhvi & Yanchong Zheng
Field Visits in Karnataka and Bihar. Image Credit: Somya Singhvi & Yanchong Zheng
The research method of this project visualized. Image Credit: Somya Singhvi & Yanchong Zheng
Principal Investigator
Y. Karen Zheng
- George M. Bunker Professor
- Associate Professor of Operations Management
- Sloan School of Management
Yanchong (Karen) Zheng is the George M. Bunker Professor and an Associate Professor of Operations Management at the MIT Sloan School of Management. Her recent research focuses on two general topics: (I) the design of incentives, technologies, and behavioral interventions to enhance efficiency, welfare, and sustainability in food and agriculture systems, with a focus on smallholder supply chains; and (II) the role of information transparency in driving environmentally and socially responsible behaviors. In her research, Zheng employs a behavior-centric, data-driven, field-based approach, and she collaborates with both public and private partners on the ground to create positive impacts to society.
Challenge:
Digital platforms have arisen as a key intervention to help improve the livelihood of millions of smallholder farmers around the world. How can we utilize these platforms to develop practical, data-driven decision support tools to improve farmers' decisions and their welfare?
Research Strategy
- Develop AI-driven market intelligence tools based on a predict-then-optimize approach to enhance farmers’ market selection decisions
- Leverage large-scale transaction and product quality data from digital platforms to prescribe recommendations for farmers' quality investment decisions
- Collaborate with organizations on the ground to devise and implement practical solutions to generate positive impacts for smallholder farmers
Project description
Millions of smallholder farmers around the world persistently struggle with low productivity and high poverty. To improve the welfare of these farmers, many countries and organizations have invested heavily in digital or mobile platforms to improve farmers’ market and information access. However, two significant gaps persist and prevent the full benefit of these platforms from being realized. First, while these platforms enable large-scale data collection, there have been insufficient efforts in translating this into data-driven decision support tools for farmers. Second, the data collected through these platforms are often “censored” in a non-random fashion. Hence, advanced machine learning and statistical methods must be developed to generate unbiased prediction and prescription. In this project, Karen, together with her collaborators Retsef Levi (MIT Sloan) and Somya Singhvi (USC Marshall), seek to fill these gaps by developing advanced analytics and optimization tools to enhance farmers’ quality investment and market selection decisions. The work will build on close collaborations with organizations on the ground to ensure that the research findings will be fully transferred to actual practice and bring material benefits to the farmers.
Outcomes
- Developed AI-driven, predict-then-optimize market intelligence tools to enhance market selection decisions of smallholder farmers by effectively predicting one-day-ahead market prices for agricultural products and optimizing market selection
- Collaborated with a nonprofit to test the application of the developed tools in a specific use case for fresh produce farmers in Bihar, India
- Developed a set of models to investigate the impact of different farming contracts on smallholder farmers' production and the resulting welfare of the farmers and the buying company
- Used field data from pork producers in China to validate the model predictions and are in the process of disseminating the insights to agricultural companies
Additional Details
Impact Areas
- Food
Research Themes
- Economics, Policy, & Practices
- Transforming Food Systems
- Modeling & Data Analytics
- Equity & Access
Year Funded
- 2021
Grant Type
- Seed Grant
Status
- Completed