Social science: Mobile phone data may improve targeting of humanitarian aid

 
Machine learning algorithms trained on mobile phone data may help recognize patterns of poverty and prioritize assistance to the poorest.A paper clarifying this will be published in Nature.In Togo, Africa, this method is used to provide millions of dollars in support for Coronavirus Disease (COVID-19) to those who appear to be most in need, with greater efficiency than traditional distribution methods. I was able to distribute it.

Following the COVID-19 crisis, governments and humanitarian organizations around the world are providing social support to more than 15 billion people.However, the challenge of quickly identifying and targeting those in need of assistance has not been resolved.Now, Joshua Blumenstock and colleagues have developed, implemented, and validated ways to tackle this challenge using machine learning algorithms that can measure poverty.Togo's most important emergency social support program came into force in April 19, shortly after the first COVID-2020 patient was discovered, and distributed money using these machine learning algorithms.

Blumenstock et al.'S artificial intelligence method is erroneously excluded (needs help, but the government's traditional support) compared to other geographic targeting methods considered by the Togolese government. The percentage of people excluded from the gold distribution program) decreased by 4-21%.In contrast, assuming that Togo has a social register (actually Togo does not have a social register), those who are erroneously excluded when compared to other methods of using it. Increased by 9-35%.

Blumenstock et al. State that the results demonstrate the capabilities of these methods when implemented on a scale in real-world crisis scenarios.

 
* This article is reprinted from "Nature Japan Featured Highlights".
Reprinted from: "Sociology: Use mobile phone data to improve the accuracy of narrowing down the target of humanitarian assistance'