Using data science to combat poverty
India is one of the world's largest food producers. However, insufficient refrigeration possibilities and other bottlenecks in the supply chain result in a waste of up to a third of the food produced – an estimated billion-dollar loss. Only 6% of the food moves through the cold chain, compared to about 60% in developed countries. The situation is especially problematic for smallholder farmers, which are major contributors to food production in India. Financial, technological, and knowledge barriers prevent these farmers from accessing sustainable cooling solutions to save food – and secure their livelihoods.
To tackle this challenge, Empa and BASE are developing an open-access, data-science based mobile app to give smallholders access to sustainable cooling facilities, pre- and post-harvest expertise, and market intelligence. The app, Your Virtual Cold-Chain Assistant (YourVCCA), includes various data inputs such as weather and climate data, satellite images, geographic location data, fresh-produce yields, hygrothermal cold-storage sensor data, forecasted remaining shelf life of produce, and real-time market prices. By equipping smallholders with appropriate data, the project aims to improve food security, increase smallholder incomes, reduce food loss, and minimize the impact of food production on the global climate.
Access to cooling facilities and data
A decisive way to tackle smallholders' access to cooling facilities is by launching partnerships with cooling solution providers. One of the partners provides cooling services in a market visited by some 500 smallholder farmers. This project, which serves as a pilot for YourVCCA, will collect hygrothermal and usage data and provide insights on optimal storage conditions and the expected shelf life of crops. The project team aims to collect datasets and consolidate them into a multi-layer map of India, displaying the most promising cold room storage sites in the country.
At the same time, the effectiveness of YourVCCA will be tested, including an image-based machine-learning model to grade fruits after harvest. Moreover, the project team is developing visual, physics-based food twins that predict the expected shelf life of the stored fruit in real time. "At Empa, we actually predict how long you can store a product and how smallholder farmers should store these products in order for them to last as long as possible", says Empa researcher Thijs Defraeye. These food twins are about to be made available to farmers and cooling service providers via a web-based platform or via an executable version that can run on mobile devices. In YourVCCA, these food twins and their shelf life predictions are combined with market price forecasts to instruct farmers when to best sell their produce. "We have designed YourVCCA as a kind of virtual coach that advises users how to store their products, how much of the quality is still there, and when they should sell", adds Defraeye.
Currently, the project team is working on a preliminary architecture of the user interface for YourVCCA. Its user friendliness for both smallholder farmers and cooling system providers will be tested in the coming months. "The progress is very promising", claims Defraeye. "If we can harness the data, we can save much more food, and this would benefit both the local communities and the climate. We all know that the technology is there, but currently the solutions are not tailored to serve smallholder farmers, as the financial incentive to do so is just lower than with larger agricultural companies."
