• Fields of the World (FOW), an open access dataset containing digital twins of international fields, has been launched on the Agrifood Data Marketplace
• FOW will enable digital solutions to be created and tested more rapidly and cost effectively.
• Agriculture, food supply chain, green finance, insurance, commodities, and academic sectors stand to benefit.
• The effort has been supported by Microsoft’s AI for Earth initiative
Data is holding back data-driven innovation. Despite there being more data than ever before, finding and accessing the right data is difficult. Increasingly, datasets need to be combined; integration is relatively simple if the data sources are interoperable, but they rarely are.
Agrimetrics, the Agrifood Data Marketplace, addresses this challenge by making data interoperable. Recently, they announced the release of Fields of the World, an open-access dataset created with the support of Microsoft’s AI for Earth initiative.
Dr. Matthew Smith, Agrimetrics’s Chief Product Officer explained:
“Agrimetrics and Microsoft saw a great opportunity to use the capabilities of the Agrimetrics Data Marketplace to make it easier for innovators to get the data they need by interoperating agricultural field data from different sources. Now, with a single query, they can pull weather, location, soil and satellite derived property data directly into their workflows, saving time and enabling them to focus more on innovating their solutions.”
Fields of the World is an open access (free to use) dataset consisting of field digital twins for 12 crops spanning 11 countries. It includes field boundaries, soil and weather data, and crop cover. FOW comprises more than 3,500 fields and over 50,000 hectares. You can find the full details here. The initiative has benefitted from the generous provision of crop-field location data from a range of different international organisations.
Fields of the World is available via API through Agrimetrics’s Agrifood Data Marketplace. API (Application Programming Interface) provides the most convenient access mechanism for analysts, data scientists and developers; the key target users. The data have been pre-sourced and pre-linked, enabling users to download and use right away. Users can write queries to identify and obtain just the data elements that they need, effectively assembling a single smaller bespoke dataset from a range of much larger datasets.
Agrimetrics’s Query Builder can be used to help users get started, providing a graphical frontend that helps with query construction and providing example queries, such as fetching all the field boundaries, cropping history, soil, weather and satellite data for all Fields of the World fields in Mexico.
The dataset has broad applications, but will be of most value to users in the early stages of hypothesis testing and product development. Fields of the World will provide the linked-data to enable creators to efficiently test their solutions without needing to invest significant time and budget.
Yield forecasting, crop anomaly detection, and optimised water management are promising early use cases. Green finance, commodities, and crop insurers stand to benefit, alongside arable agriculture.
“There is a huge untapped potential from applying AI to interoperable datasets”, says Matthew. “Innovators around the world are discovering how to derive new information from applying AI. For example, our own Data Science team have discovered new ways to forecast crop development and identify crop species from satellite data. This is what is needed to make remote monitoring and validation cost-effective and sustainable and this is being sought from businesses throughout the Agrifood system, especially those involved in ensuring continuity of food supply from the farm to the consumer, including financial organisations and retailers.”
Agrimetrics is a Microsoft Gold Partner and graduate of Microsoft’s AI for Earth programme. Microsoft has supported creation of the Fields of the World dataset through an AI for Earth grant as part of their mission to accelerate those working to solve global environmental challenges.
Dan Morris, principal scientist, Microsoft AI for Earth commented:
“Enabling data-driven agriculture at a global scale is critical to both climate change resilience and minimising human impact on ecosystems. The Fields of the World dataset is an example of how global-scale data and cloud-scale computation can be combined to provide field-level decision support information that previously would have required redundant, expensive effort across thousands of individual field operators.”
To help illustrate one application of the Fields of the World data, Agrimetrics’s team of Data Scientists created an example use of the data to drive a crop growth model for wheat, applying their model to the UK, France and The Netherlands. The model used the Fields of the World weather and soil data as inputs and simulates the development of crop growth over time, resulting in predictions of crop yield. Technical readers can explore the Jupyter notebooks, here.