Food insecurity is a major concern affecting people across the world. According to the Food and Agriculture Organization of the United Nations, the number of undernourished people increased from 624 million in 2014 to 688 million in 2019. This is a worrying trend, and the situation has only gotten worse due to various factors such as the COVID-19 pandemic, climate change, and armed conflicts. In 2021, between 702 and 828 million people worldwide faced hunger, and severe food insecurity increased both globally and in every region.
Despite the gravity of this issue, current methods to detect future food crises rely on risk measures that are insufficient, making it difficult to address them. In response to this, researchers sought to develop a better model that could help prioritize the allocation of emergency food assistance in vulnerable regions, surpassing current methods.
Samuel Fraiberger, a visiting researcher at New York University’s, said, “Our approach could drastically improve the prediction of food crisis outbreaks up to 12 months ahead of time using both real-time news streams and a predictive model that is simple to interpret.”
Lakshminarayanan Subramanian, a professor at the Courant Institute, said, “Traditional measurements of food insecurity risk factors, such as conflict severity indices or changes in food prices, are often incomplete, delayed, or outdated.” and added “Our approach takes advantage of the fact that risk factors triggering a food crisis are mentioned in the news prior to being observable with traditional measurements.”
The researchers considered the possibility that news coverage, which offers real-time, on-the-ground accounts of local developments, could serve as an early-warning system for impending food crises. They collected text from over 11 million news articles focused on nearly 40 food-insecure countries published between 1980 and 2020. They then developed a method to extract specific phrases related to food insecurity from the articles and in ways that capture journalistic assessment in notable detail.
The tool used by the researchers accounted for nearly 170 text features to correctly gauge the semantics of the phrases related to food insecurity and to mark when the articles appeared. An example from South Sudan outlines both the location and risk factors, stating that “famine may return to some parts of the country, with eastern Pibor county, where floods and pests have ravaged crops, at particular risk.”
The researchers then considered data on a range of food-insecurity risk factors such as conflict fatality counts, rainfall, vegetation, and changes in food prices to determine if there was a correlation between news mentions of these factors and their occurrence in the studied countries and regions. Here, they found a high correlation between the nature of the coverage and the on-the-ground occurrences of these factors, indicating that news stories are an accurate indicator of the studied conditions.
To determine if news articles were a good predictor of subsequent food crises, the researchers needed to know if the nature of the coverage was a viable indicator of future crises and if these stories did so more accurately than traditional measurements. Using a smaller set of news stories, the researchers found that news coverage yielded more accurate predictions at the local level of food insecurity and did so up to 12 months ahead of time than traditional measurements that did not include news story text.
Moreover, supplementing traditional predictive measures with news coverage further improved the accuracy of food-crisis predictions, suggesting the value of “hybrid” models. The researchers see potential for larger applications of their work and hope that it can help prioritize the allocation of emergency food assistance in vulnerable regions, ultimately helping to alleviate food insecurity for millions of people worldwide.
Study Credit: https://www.science.org/doi/10.1126/sciadv.abm3449
