U.S – Researchers at the Central European University’s Department of Network and Data Science have developed a new tool that can use a variety of data to anticipate food insecurity up to 30 days in advance.
Elisa Omodei, a study author and Assistant Professor in the Central European University’s Department of Network and Data Science, claimed that the tool could help decision-makers in nations at risk of food insecurity by facilitating more prompt responses.
The 2030 Agenda for Sustainable Development, adopted by all United Nations Member States in 2015, calls for urgent action to “end hunger, achieve food security and improved nutrition and promote sustainable agriculture”.
However, in 2019, 650 million people were still undernourished, with 135 million in 55 countries and territories reported to be acutely food insecure, say the researchers.
These numbers have significantly increased as a consequence of the COVID-19 pandemic, with at least 280 million people reported to be acutely food insecure in 2020.
The researchers examined data from 2018 to 2022 using food consumption statistics from countries that have recently experienced food insecurity like Burkina Faso, Cameroon, Mali, Nigeria, Syria, and Yemen.
The authors then added information about deaths related to conflicts, food costs, bad weather, and the presence of Ramadan during this time to their tool.
The percentage of households at risk of having insufficient access to food between October 2021 and February 2022 was then estimated using the method.
According to their research, the technology was able to predict the frequency of food insecurity in Yemen and Syria with 99 percent accuracy for one day and 72 percent and 47 percent accuracy for 30 days, respectively.
However, they acknowledged that the tool’s predictions were not as precise when it came to the data for the four remaining countries. This, the researchers said, is because there was less accessible data on food consumption in Syria and Yemen.
“This highlights that the tool’s forecasts are more accurate when using food consumption data collected at regular intervals over long periods and across a broad range of geographic areas,” said the researchers.
In Cameroon, only a few regions were characterized by a relatively high proportion of food insecure people, generally above 50%, but also exhibiting large fluctuations, such as the rapid decline and subsequent increase observed in the North-West regions.
On the other hand, in Syria, the sub-national trends were all similar in terms of relative changes in the affected population, with a general upward trend affecting almost every province beginning in July 2020.
In the governorates of Yemen, for which the longest time series are available, the proportion of the population affected by food insecurity varied between 20% and 60% during the years 2018–2022, however, a common national time trend was less recognizable.
“It should be noted that some of these irregularities could also be partially due to the effects of sampling and selection bias, whose mitigation can rarely be achieved in full,” the authors point out.
The authors explain how the tool can be helpful by saying that it could supplement current methods for modeling food insecurity by offering quickly accessible forecasts using real-time data.
According to them, future predictions made using the technique could be enhanced by the inclusion of mobile phone data or automatically generated news text mining.