New tool predicts food insecurity up to 30 days in advance

April 15, 2023 | Food Testing Lab

Elisa Omodei, principal author of the study and an Assistant Professor Central European University- Department of Network and Data Science, claimed that a tool can forecast the prevalence of food insecurity up to 30 days in advance using various data. This could aid and facilitate timely responses especially for decision makers in countries at risk for insufficient food access.

Omodei and team used and analyzed food consumption data within 2018-2022 from countries that have recently experienced food insecurity such as Burkina Faso, Cameroon, Mali, Nigeria, Syria and Yemen.

They further enhanced their tool by including data on conflict-related fatalities, food prices, extreme weather events and the occurrence of Ramadan among others. The tool was able to estimate the prevalence of households at risk of food insecurity from October 2021 to February 2022.

Results from the tool were able to forecast (with 99 percent accuracy one day into the future; 72 percent and 47 percent respectively 30 days into the future) regarding the prevalence of insufficient food access within Yemen and Syria. Because of less available food consumption data of the four remaining countries, the tool’s predictions were not as accurate. This meant that the tool’s forecasts have greater accuracy if food consumption data collected at regular intervals over long periods of time across a broad range of geographic areas.

The researchers stated that their tool would be able to complement existing techniques that model food insecurity by rapidly providing forecasts using real-time data. Furthermore, the tool’s prediction can be enhanced by involving mobile phone data or by automated text mining of news.