Technical Report on the Turba Fertilizer Machine Learning Stack in Morocco
Turba’s open-source stack packages reproducible site-specific fertilizer recommendation models for Moroccan cereals.
The technical report describes Turba, a three-layer open-source stack for reproducible site-specific fertilizer recommendations in Morocco. turba-client accesses site profiles and nitrogen, phosphorus, and potassium workflows; turba-data ships versioned snapshots; and turba-models packages crop-specific surrogates. The first dataset covers 44,096 ESA WorldCereal locations and 132,017 medium-yield requests across 10 regions, 66 provinces, and 1,149 communes. Nine regression families were compared under a fixed 80/20 split, and five best crop-specific models are released. The task emulates recommendation functions rather than predicting observed crop response.
- Three layers: turba-client, turba-data, and turba-models.
- Built from 44,096 locations expanded to 132,017 recommendation requests.
- Covers 10 regions, 66 provinces, and 1,149 communes in Morocco.
- Five best of nine regression families are packaged for offline use.
- Models emulate recommendation outputs rather than observed crop yields.
From the original article
1 min read at huggingface.co
Site-specific fertilizer recommendation systems adapt nutrient advice to location, soil properties, crop type, and production targets, but scientific reuse is constrained when recommendation functions remain accessible mainly through interactive interfaces, outputs are not versioned, and trained approximations cannot be independently loaded or benchmarked.
By Abdelghani Belgaid, Zakaria Mahmoud, Fahd Chibani · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.