Dr. Shai Sela

The quantity and scope of agronomic data available for researchers in both industry and academia is increasing rapidly. Data sources include a myriad of different streams, such as field experiments, sensors, climatic data, socioeconomic data or remote sensing. The lack of standards and workflows frequently leads agronomic data to be fragmented and siloed, hampering collaboration efforts within research labs, university departments, or research institutes. Researchers and businesses therefore allocate significant time resources into unifying these fragmented data layers into a coherent structure. Implementing data standardization schemes can enable efficient collaboration, and leveraging the collective power of the research community to address critical agronomic knowledge gaps. This presentation will provide an overview of available research data standardization tools and explain the underlying FAIR and other data management principles. Using Agmatix’s Axiom platform as an example, we will demonstrate how data from multiple sources can be standardized and used for insightful modeling. We have used 3774 experimental data points from multiple sources in the United States – universities, commercial associations and farm-management systems – to construct a corn prediction model (R2 = 0.9. RMSE = 1.0 Mg/Ha). The standardization platform unifies all datasets to common headers and units, allowing to explore data distributions to ensure an adequate crop yield parameter space is covered. While demonstrated here on corn yield, many different agronomic research domains can benefit from standardization of data. We call on the agronomic research community to adopt standardization, boost collaboration efforts, and increase the potential for tackling the current global agronomic challenges.