Forecasting Famine and Food Crises

Famines and related food, nutrition, and health crises lead to extreme human suffering, enormous financial costs, and potential instability.

Recent advances in theory, modeling capability, and real-time digital data collection have the potential to improve our understanding of these crises and our ability to predict and respond to them in a timely manner. Yet there are real obstacles. For example, in resource-constrained environments, and during crises, data on deaths, injuries, diseases, disability, mental health conditions, and malnutrition are difficult to collect. In addition, political sensitivities and agendas often conflict with humanitarian priorities and combine with uncertainties about forecasts to create barriers to action.

Given the complexity of historic and current crises, we must use a multi-disciplinary approach to understand, predict, and respond to food crises in a transformational manner. We believe that “Precision Humanitarian Response” is required. Therefore, this project is harnessing the power of interdisciplinary teams and community participation to co-create solutions that will guide better tailored responses to food crises. Our technology hub has two main components:

  • Famine Archive. We are developing an inventory of data sources and their curators, for historic and current famines and humanitarian crises.
  • Hackathons. We run Hackathons to examine the potential use of the information in the Famine Archive. During the Hackatons, we use innovations in information technology, AI, computational capacities, and novel data sources to develop modeling suites and build interactive dashboards and visualizations.

This project draws on theoretical and practice-oriented work at the Feinstein International Center and modeling assets of the Tufts Initiative for the Forecasting and Modeling of Infectious Diseases (InForMID).

Our Hackathons bring together a broad interdisciplinary community of researchers and practitioners to:

  • Foster a learning community of enthusiasts in humanitarian and data science fields
  • Facilitate the exchange of ideas across different disciplines, from data science to theorists to humanitarian practitioners
  • Learn to work effectively in multi-generational groups with varying levels of technical skills and expertise
  • Understand the theory and concepts of famine formation, evolution, and resolution
  • Sharpen statistical, computing, and visualization skills
  • Recognize the challenges of both humanitarian action and data science
  • Understand limitations and opportunities for forecasting in humanitarian contexts
  • Deepen the dialogue around data quality, availability, accessibility, and completeness and implications for bias, misinformation, distortion, and mistrust
  • Develop and curate datasets to support future research

Recent Hackathons

Our Hackathons use data from the UNICEF Nutrition Cluster in Yemen, which contains monthly measures of various nutrition indicators.

During the first Hackathon (February 2023), three teams explored the data and created dynamic maps.

During the second Hackaton (March 2023), four teams developed, designed, and deployed impactful visualizations (dynamic maps, dashboards, and posters) to raise awareness and promote action on famine in Yemen.

Our third Hackathon (upcoming date to be determined) will build on the work of the first two events.