open-turba/turba-data

Dataset package for soil fertility and fertilizer recommendation workflows across Morocco. turba-data Dataset package for site-specific NPK fertilizer recommendations in Morocco. `turba-data` is the data package of the turba ecosystem. It provides a simple way to discover the datasets currently available and load them directly into Python. Future work will expand the catalog with additional datasets covering more crops, sources, and regions, add more notebooks for analysis and comparison, and introduce hosted remote downloads as the collection grows. ## Install ```bash pip install turba-data ``` This also pulls: - `pandas` - `pyarrow` ## Quick start ```python import turba_data as td print(td.list_datasets()) df = td.load_dataset("esa_worldcereal_morocco_cereals_medium") print(df.head()) ``` Users can convert parquet to CSV themselves after loading: ```python import turba_data as td df = td.load_dataset("esa_worldcereal_morocco_cereals_medium") df.to_csv("esa_worldcereal_morocco_cereals_medium.csv", index=False) ``` ## First release dataset Current snapshot for the first packaged dataset: - dataset id: `esa_worldcereal_morocco_cereals_medium` - file: `esa_worldcereal_morocco_cereals_medium.parquet` - rows: 132,017 - columns: 22 - unique sites: 44,096 - regions: 10 - provinces: 66 - communes: 1,149 ## Repository structure ```text turba-data/ ├── src/ │ └── turba_data/ │ ├── __init__.py │ ├── core.py │ ├── registry.json │ └── datasets/ │ └── esa_worldcereal_morocco_cereals_medium.parquet ├── tests/ │ └── test_api.py ├── LICENSE ├── pyproject.toml └── README.md ```