The Future of Data-Driven Agriculture

Last week USAID and the Sustainability Innovation Lab at CU-Boulder (SILC) hosted a two-day interactive workshop on leveraging big data to further agricultural development around the globe. The convening, held in Boulder, gathered a cross-disciplinary group including researchers, agronomists, digital agtech startups, data scientists, non-profits, foundations and venture capital firms to examine how data science can maximize impact in agricultural development.

Danielle Dhillon from DIAL speaks about the Data for Development team at USAID Innovation for Data-Driven Agriculture Workshop on April 27, 2017.

During the event, several key themes emerged including the importance of smallholder engagement particularly to keep stakeholders at the front of planning, the role of identifying and leveraging the best communication channels to deliver information and gather feedback, and the importance of delivering the right insights at the right time.

Using data and digital services for development in the agriculture sector can make a big impact from collecting and analyzing real-time data and delivering actionable insights. During this event, attendees heard from a couple players in the agtech space who are using data-driven approaches to tackle some of the toughest challenges in the agriculture sector.

  • Land-Potential Knowledge System (LandPKS) is a program that develops innovative web-based and mobile data collection/analysis methods and tools to support local land use planning and to optimize design and implementation of food security, land restoration, climate change adaptation, and biodiversity conservation programs. Currently they offer a suite of free mobile and web applications that serve as site characterization and vegetation monitoring tools for use by anyone. At the workshop they presented SoilID, an app that can be used for soil identification by combining simple user inputs and traditional soil map information. This app offers a cheaper and faster alternative to traditional soil testing. By having a user interface that directly interacts with farmers, LandPKS can dilute data from soil maps into digestible insights as well as collect data that could be pertinent to a nearby farmer. This app is an example of how digital services can serve as a conduit to bring data science out to the field.
  • aWhere operates a global-scale agronomic modeling environment to deliver agricultural intelligence into the hands of farmers, commercial growers, commodity traders and policy makers. They provide weather and agronomic data for timely decision making in areas including but not limited to farmer productivity, academic agriculture and food security. The aWhere platform is built around localized weather information utilizing over 1.5 million weather stations and hyper-local coverage. They can combine this data with moisture stress and crop stress indices, allowing farmers to adapt locally to weather, pest, and disease forecasting. The applications for this data are vast; it can be tailored and disseminated through SMS as packaged agronomic tips or weather alerts. It can also be used as a backbone to weather-indexed insurance. Also, when combined with crop-specific and field-specific agronomic information, it can generate valuable analyses to help farmers reduce risk and increase crop yields.
  • Digital Green is a not-for-profit international development organization that uses an innovative digital platform for community engagement to improve lives of rural communities across South Asia and Sub-Saharan Africa. Through unique channels such as locally produced videos and human mediated dissemination, they share knowledge to improve agricultural practices, livelihoods, health and nutrition. They use a data management software customized to low resource settings to collect and analyze near real-time data on dissemination, adoption, and community interest. The majority of this data requires the manual input from their extension workers because some of the data is qualitative (e.g. feedback and adoption practices) and/or requires a physical person to count (e.g. average attendance per dissemination). What is unique about Digital Green is that it has the ability to accept data while offline or areas with intermittent internet connectivity. They host their analytics dashboard online through their free and open-source COCO database, which they make available to others. In the data for development space, Digital Green is fascinating because of their use of digital services to deliver information to off-grid locations as well as their ability to capture unstructured data to a platform even when they are offline.

Danielle Dhillon joined the Digital Impact Alliance in March 2017 as Senior Program Analyst, Data for Development. In this role, she works to demonstrate the value of a viable Data for Development (D4D) ecosystem for driving effective learning and decision-making across development programs, the public sector and the private sector.