Radiant Earth Foundation

I joined Radiant Earth Foundation in 2017 as a Senior Data Scientist and was mandated to build the Machine Learning portfolio for the organization. Since then and as a result of what I started, we have redefined our mission, narrowed down our focus to enabling an ecosystem, and I have transitioned to the leadership role. 

The funding from the Elevate Prize will help me lead our team in taking Radiant Earth to the next level. While we have been fortunate to receive project funding from several organizations, this unrestricted funding would help us expand our infrastructure and prime us to scale our impact. 

The mentorship from the Elevate Prize community will be an invaluable resource to me and our organization. I am eager to get connected to like-minded leaders who use technology for positive impact, and learn from their experience to be able to lead Radiant Earth forward. The networking and marketing support will also help us reach new potential funders who would be interested in supporting our mission, as well as users and communities who can benefit from our service or contribute to its growth. This prize will give us the opportunity to have a bigger impact!

We live in an era where many technologies are rapidly evolving and providing opportunities to address international development challenges. I believe, however, that we will see an increasing digital divide unless we empower individuals to access these technologies to build local solutions to their own local problems. 

Earth Observations (EO) which provides invaluable data for deriving insights in agriculture, climate, clean water, energy, etc, combined with Artificial Intelligence (AI), can be a game-changer in how policy makers use data to make decisions. I envision that these two technologies will reduce the need for ground-based data collection infrastructure, and enable various stakeholders from public, academic and commercial sectors to map regional or national resources. 

That is why I am excited to lead Radiant Earth Foundation on a mission to make decision making from EO data more efficient, lower the barrier to entry for those who want to use these technologies to deploy innovative solutions, and address the gap and bias in related datasets and models. 

Our goal is to establish an ecosystem of users, data, and knowledge. We do this through an open-access technology hub, convening the community, and organizing training and data science competitions.

The World Bank reports that only 57% of African countries have carried out an agricultural survey during 2007-2017. Meanwhile, agriculture constitutes at least 15% of the GDP in 60% of African countries. Lack of such data about a crucial segment of the economy impacts the lives of millions of people across the developing world and impedes their progress towards SDGs of No Poverty, Zero Hunger, and generally sustainable resource management. 

We aim to solve this problem (and similar ones that benefit from EO data) by empowering local communities to build solutions using EO and AI through three pillars of activity:

1. Radiant MLHub: The first open-access repository for benchmark geospatial data/models. 

2. Communities of practice: Convening domain experts and data scientists from around the world to develop guidelines and standards on how to use EO and AI technologies effectively and more responsibly.

3. Training and education: Providing tutorials, organizing boot camps, and sharing the latest advancements of the field with our user community.

Through democratizing access to EO and AI, we equip practitioners and decision makers with the information and tools they need to develop practical solutions to social, economic, and environmental challenges at the local, regional and collectively, global level. 

Ion Stoica, Professor at UC Berkeley once said: “Data is only as valuable as the decisions it enables.” I believe for data to enable decisions, we need data-collaborative innovation. 

Data has a multi-stage value chain (see figure below). It starts with Collection, then Publication, next to Uptake, and finally Impact. Our work aims to address gaps in the Publication stage, at the same time, ensure integration and standardization across other stages. While our work is focused on the Publication stage, we do work on better tooling and standards for Collection and enhanced pipelines to facilitate Uptake. 

Our innovative approach is that we act as a neutral entity working cross-sectoral with stakeholders/users who will benefit from better access to the data, as well as data providers who would like to serve those users. We work with and actively engage government, academic, commercial and nonprofit organizations to enable a smooth flow for the data from Collection to Impact. Lastly, our standards and guidelines are defined with the users, for the users, and this has been our success in developing specifications such as SpatioTemporal Asset Catalog (STAC) that is now being widely used across the geospatial sector.  

Good data is crucial for the global development community to prosper. Public and private sector decision makers regularly use data to make decisions, whether to allocate resources according to the population or economic statistics, or to identify opportunities and attract new investments. 

Nevertheless, many developing countries do not have the infrastructure or organizational establishment to deploy people on the ground to collect data. Establishing a ground-based data collection infrastructure is costly, and inefficient. These countries can leapfrog and benefit from the technological advancements of EO and AI to more efficiently collect their required data and insights. For this to happen, we need to democratize access to EO and AI. This is what Radiant Earth Foundation is determined to do.

We currently host two dozen benchmark datasets for various applications including agricultural monitoring in several countries including Kenya, Tanzania, Uganda, Benin, and Rwanda. As of May 2021, we have +2100 users accessing Radiant MLHub. In 2020 alone, we had more than 1000 participants in our online workshops and training events. In addition, 666 participants competed in our data science competitions, with top awardees from Algeria, Egypt, Tanzania, Tunisia, Nigeria, Kenya, Uganda, Peru, and the United States.