Logo Lanfrica
  • Accueil
  • Atlas
  • Analyses
  • Documentation
  • Sign in

© 2026 Lanfrica. Tous droits réservés. Tous les droits d'auteur des ressources affichées sur le site Web Lanfrica appartiennent aux détenteurs de droits d'auteur d'origine, sauf indication contraire explicite.

Valerie-komen/Maji-Ndogo-Farm-Yield-Prediction

Domaine:

agriculture

Type de record:

project
Créateur:
Val
Hôte:
Regression modeling of crop yield per hectare from geographic/weather/soil data, with two real data-quality bugs found and fixed before modeling. # 🌾 Maji Ndogo Farm Yield Prediction **Regression modeling of crop yield per hectare from geographic, weather, and soil data — with two real data-quality bugs found and fixed before any model was trained** > Interactive dashboard: dashboard/farm_yield_dashboard.html --- ## Problem 5,654 agricultural fields across 5 provinces, 8 crop types, with geographic, weather, soil, and farm-management data. The obvious modeling target — a column literally named `Annual_yield` — turns out to be the wrong one to predict (see below). This project finds the right target, engineers a clean feature set, and builds a yield-prediction model that explains which environmental factors actually drive yield. **Research questions:** - What does "yield" actually mean in this dataset, and is the most obviously-named column the right one to model? - Which environmental factors (soil, weather, pollution, geography) correlate with yield? - Can yield per hectare be predicted from field characteristics, and how well? --- ## Dataset Source: `Maji_Ndogo_farm_survey_small.db` (4 linked tables: geographic, weather, soil/crop, farm-management features — 5,654 fields) + `Weather_Station_Dataset.zip` (1,843 free-text weather-station messages). --- ## Two real bugs found before modeling 1. **`Crop_type` and `Annual_yield` are swapped in the source schema.** The column labeled `Crop_type` held continuous numbers; the column labeled `Annual_yield` held crop-name strings (with typos: "wheatn", "teaa", "cassaval"). Verified and fixed in `notebooks/01_data_cleaning.ipynb`. 2. **`Annual_yield` is not a usable modeling target.** It correlates at r=0.95 with `Plot_size` — it's just total farm output, which scales with farm size almost by definition. The `Standard_yield` column, despite its name suggesting a fixed benchmark, turns out to be **exactly** `Annual_yield / Plot_size` (r=1.0, verified in `notebooks/02_exploratory_analysis.ipynb`) — the real, size-independent yield-per-hectare rate, and the c …

Visit

github.com

Languages

DizinNdogo

Similaires

Valerie-komen/Maji-Ndogo-Water-Access-Analyticsikigai-hub/maji-ndogo-farm-analysisutiwisdom/maji-ndogo-crop-yield-analysis

Valerie-komen/Maji-Ndogo-Water-Access-Analytics

SQL + Python analysis of water infrastructure equity across 27.6M people: data reconciliation, EDA,

ikigai-hub/maji-ndogo-farm-analysis

What factors influence crop yield across 5,654 farms in the Maji Ndogo region, and why do some farms

utiwisdom/maji-ndogo-crop-yield-analysis

Analysis of factors affecting crop yields in Maji Ndogo using regression techniques. This project ex