Data Collection and Publishing
Why was the project undertaken?
For a long time, many stakeholders in the Agricultural Sector have found it hard to accurately quantify the area under production and yield particularly, which reduces visibility in the sector. This makes it hard for financial institutions to commit themselves in funding agricultural activities and also for insurance companies to insure crops.

The georeferenced yield data collected will help to generate yield maps through use of Remote sensing and machine learning and also create platforms for monitoring crop health for farms throughout the crop life cycle. Yield estimation will also help crop insurtech companies have more data for pricing and financial institutions can also have enough data that will give them confidence to fund agricultural activities.
How was the data collected?
Data was collected using FAO’s, Crop Cutting Experiment involving either 2 or 3-step protocol depending on the crop. The 3-step protocol (visits) was done on all crops except Irish potatoes.

During the first visit, a box of 8 meters by 5 meters was laid for crops other than Irish Potatoes, while a square box of 5 meters was used for Irish potatoes. The boxes were placed just after the crops flowered to prevent farmers from harvesting.

During the second visit, the crops are harvested, weighed and stored in sacks to ensure they lose some moisture for the final visit, except Irish Potatoes whose wet weight is the final weight.

During the final visit, the crops were threshed and the final dry weights were captured.

The weight captured at this stage is considered the final weight of the crop and the average yield MT/Ha is calculated depending on the size of the box placed.
Where is the data collected from?
Labelled and georeferenced yield data were collected from two countries, Zambia and Kenya.

Crop types: The crop types include Irish Potatoes, Maize, Green Grams, Beans and Sorghum.