Part of UNICEF’s data for social good initiatives, UNICEF Innovation is building Magic Box, a software platform designed to use real-time data to inform humanitarian responses to emergency situations. SIX spoke to UNICEF’ Innovation’s Clara Palau, Technology Team Lead, and Manuel Garcia-Herranz, Chief Scientist.
UNICEF Innovation developed Magic Box in response to emergencies, first during the 2014 Ebola crisis in West Africa. In collaboration with Google, the platform was updated in response to the Zika outbreak in 2015. This open-source platform takes in data from both public sources and private sector partners to generate key insights using different methodologies and algorithms provided by UNICEF’s data science team. Magic Box is being utilized to fight epidemics, map schools, track sudden onset disasters and recovery, and measure social indicators of poverty. Current partners span the public and private sector and include Amadeus, Telefonica, Google, IBM, Facebook, and academia.
Why was Magic Box created?
When the Ebola crisis became an emergency in 2014, UNICEF began using data to analyse how the disease spread. However, establishing the partnerships UNICEF needed and building the infrastructure to obtain and use this data took months, which was too long to respond to a humanitarian emergency. This highlighted the need for a platform to develop key partnerships and methods before an emergency struck.
UNICEF partnered with Google, as well as academia and data scientists, to create the infrastructure for Magic Box. The platform is unique in that it is open source and allows for the combination of different datasets from diverse providers. While natural disasters and disease outbreaks are often unpredictable, the platform enables existing knowledge to be assembled ahead of time instead of negotiating data access reactively.
Most data is currently owned and used by the private sector and the development world has historically played catch-up in applying it to humanitarian contexts. The goal of Magic Box is to bring in new sources of data, including big data, from mobile operators, social media, travel data, and more, as well as new computational techniques like machine learning to UNICEF to help address complex humanitarian and development challenges.
What are the major challenges in using data for social good?
Data and artificial intelligence have grown rapidly in the past decade, mostly from Silicon Valley and the world’s richest markets. These methods have been developed in scenarios not always easily transferable to a humanitarian context. For example, “real-time mobility insights can be gained from mobile phone data, but often 30% of people in developing countries do not have a phone, and this is often the most vulnerable population you are trying to reach.” UNICEF is calling out to the wider community to look for ways to address this challenge.
Another connected challenge is a lack of technical capacity and data literacy within the UN and humanitarian environment, something SIX has noticed across the sector: “It is not easy to speak on equal terms with private or academic data scientists or AI experts, as there is a gap in problem understanding” – largely due to the lack of experience in using data for humanitarian contexts. There is not a lot of mobility between the sectors, between data science and the humanitarian world. “There is a need for more computer science co-working with development, for resumes with both data science and humanitarian backgrounds.”
There are also challenges with the data itself. The first is privacy. “Magic Box uses a high level of aggregation and all data is anonymized. While this could lead to a loss of some information, it is critical to protect the privacy of individuals, especially children.” Second is gaining access to the data. There is a commercial challenge, as this data has high value in the market and a cost to extract and refine, in order to give to UNICEF. There are many in the private sector moving towards data for good in the past two years, but there is still a need to make a business value case to encourage participation.
UNICEF is developing new partnerships with the private sector to access more data, particularly real-time data provided in a long term and consistent manner. They are also trying to build a larger network of modelers and to enable a community that can input new models into the Magic Box platform.
They’re using the platform in different pilot projects including Zika in Latin America with the WHO, understanding information poverty, and estimating socio-economic indicators. Scaling these projects involves working with decision makers in the field and country offices.
The work of Magic Box through its platform and partnerships reflects the wider beliefs in co-creation and collaboration to work for the most vulnerable.
As part of the SIX Funders Node, SIX is leading a new global project exploring how data can be used to help cross-sector partnerships address complex problems.
Photo credit UNICEF.
Want to read more great content like this? Subscribe to our newsletter for regular updates.