The Data for Children Collaborative with UNICEF

Using AI to develop sustainable census-independent population density estimation

National Outcome: We grow up loved, safe and respected so that we realise our full potential

The Data for Children Collaborative, a joint partnership between UNICEF, the Scottish Government and the University of Edinburgh’s Data Driven Innovation Programme, was established to investigate ways of using data to improve the lives of children around the world.

The group is running a project in Mozambique that uses sustainable machine learning models to improve current population estimates. Census data is not collected frequently enough to account for rapid population change in many emerging economies and is also found to lack precision, particularly in semi-urban and rural areas. UNICEF needs a more accurate understanding of the number of children in each community to better plan and deliver key services, such as vaccination programmes.

A multi-disciplinary team from the University of Edinburgh is using state-of-the-art machine learning tools to extract features from satellite images that are relevant to population density estimation, such as building footprints. Micro census data from UNICEF is used to train and validate the machine learning models, providing a user’s perspective on the sustainability of these tools. Ultimately, it is hoped that UNICEF will apply the optimal model identified from the project in other countries to improve their programming capabilities and provide better healthcare services to children.

By supporting UNICEF in this initiative, academics and researchers in Scotland are making a global impact by applying their expertise to solve a problem that has existed for decades.