‘Cropin’ is Using Data Lakes to Democratise Agricultural Data

Cropin's rich data lake serves as the foundation for all analytics performed on the platform. Globally, the company has computed intelligence on 0.2 billion acres of farmland. The resulting data lake is adaptable and highly configurable, allowing the company to get a 360-degree view of all agricultural inputs entering the system. This allows for the management of a wide range of crops, land, and demographics.

"The data lake gives us a single platform with a unified view of all data points." We can now retrieve data in a central and schematized manner with relative ease. Our data lake is more scalable, so it will expand to meet the system's needs as we add more assets and farmers. Using our AI/ML platform designed specifically for the Agri ecosystem, Cropin has computed 0.2 billion acres of farmland in 12 countries, spanning 24 important commodities," said Rajesh Jalan, Cropin's CTO.

Cropin has developed an artificial intelligence-enabled platform that uses satellite imagery, weather, and ground data to forecast farm performance around the world. The model has already been used at scale to predict a variety of crop parameters, such as crop detection, crop health and stress, irrigation needs, yield, and so on, allowing businesses and farmers to use the data to make proactive and predictive decisions.

With an accuracy of 85 to 95 percent, its AI models generate multiple layers of intelligence on every pixel of farmland.

"Cropin has participated in a Nigerian government project in collaboration with the Flour Milling Association of Nigeria (FMAN) and other industry stakeholders in one of the use cases. "In addition, the company used its own AI/ML and deep tech capabilities to monitor regional food security," Jalan explained.

Wheat is one of Nigeria's most important crops. Despite favourable climatic and edaphic conditions, however, wheat production in Nigeria had not yet reached its peak. The Federal Ministry of Agriculture and Rural Development had to develop a national and systematic data collection on where and how wheat is grown across the country in order to grow Nigeria's wheat value chain. Cropin used its AI/ML-powered Ag-intelligence platform to estimate wheat production.

There were two stages to the solution's development. Primary data (satellite, weather, etc.) and secondary data (land use and land cover estimation with geotagged plots in 13 Nigerian states) were collected in the first phase. The crop detection and yield model were deployed in these Nigerian states in the second phase.

Several AI/ML methods were developed, and the best model for deployment was chosen. A validation based on rigorous statistical analysis was performed to assess the yield estimates.

"This exercise was a huge success across the 13 states," according to the National Bureau of Statistics (NBS) of Nigeria's Report of Wheat Production Survey in Nigeria. He explained, "We were able to build and fine-tune a crop detection model that accurately identified wheat and wheat-producing farmlands across northern Nigeria."

CropIn's AI-based solutions use satellite images, IoT devices, data analysis, and machine learning to detect crop patterns and forecast crop futures. It provides hyper-local insights into crop health and environmental conditions in order to improve crop outcomes.

The company uses Amazon SageMaker to manage large machine learning workloads and AWS Lambda to deploy training models on serverless frameworks, allowing farmers to make better data-driven decisions. It closes the information gap and ensures that all supply chain points are aware of risks and opportunities.

Jalan also stated that the company is currently developing the world's first intelligent Ag Cloud, which will assist its B2B customers and end-users, i.e. farmers, in leveraging and reimagining the power of data in the Agri ecosystem.