Global hunger is a pressing problem. With the rise in population, income and changes in diet, ensuring households have access to affordable, nutritious food will only become more difficult in the future. Clean air, water and productive land are vital to sustain this growing food production but all of these natural resources are under stress. Data analytics are increasingly powerful but have largely not been applied to global food security and resource management because of key gaps in data availability and modeling. This initiative intends to fix this.
We face the global challenge to feed a world of 10 billion people by 2050. As incomes rise, people demand more than just calories from their food as they search for a broader range nutrients and taste. This additional demand for dietary diversity means that the 33% projected increase in population translates to a 60% increase in needed food supplies. Producing sufficient nutrients for all is made more difficult by the fact that water and soil resources are already stressed, and will be further strained by climate change.
Improving food security faces a number of data challenges. We are only now beginning to get a detailed view of how environmental inputs affect food production. This work involves downscaling of complex climate models, both in space and time and linking them to fine-grained data on agricultural production, markets and household behavior. These models of the food system further need to incorporate the risks posed by climate and political shocks. We are currently experiencing a dramatic increase in agricultural research capacity in developed countries using a combination of detailed environmental and genetic data, careful experiments using precision agriculture and complex computational models. These advances have lagged in developing countries, largely due to data constraints. While progress is underway collecting detailed physical data, social and economic data are sparse and infrequent. Further, spatially and temporally detailed data on food security outcomes are rare. Increased agricultural productivity does not always translate into improved household food security, and the full set of links between these two outcomes have yet to be understood. Big data analytics are beginning to allow us to better identify the relation between income shocks, food access and local food availability in developed countries and these tools could be extended to developing country settings. These questions not only require improved data, better data analytics, but cooperation and insights from multiple disciplines working across developed and developing country settings.
This initiative has the goal of bringing researchers from multiple disciplines together to use big data techniques to help families around the world have access to affordable and nutritious food. We will enable the collection and analysis of granular and temporally detailed data on weather, production, consumer and producer behavior and market outcomes to better understand and target food insecurity. The improved ability to develop detailed models of these complex systems and ‘big data’ techniques such as machine learning can enable the use of these detailed data to better predict food security outcomes, increase agricultural production and guide the interventions needed to reduce current and future world hunger.