Optimizing maize yield using hierarchical modeling, machine learning, and geospatial analytics
Optimizing Nitrogen Application Strategies
for Maize Yield in Rwanda
## Analysis Charter:
Optimizing Nitrogen Application Strategies for Maize Yield in Rwanda: A Hierarchical Modelling and Geopspatial Analytics Approach.
## Overview
This project investigates fertilizer application strategies and their impact on maize yield across different agroecological environments in Rwanda.
---
## Repository Structure
```text
├── dashboard/
|
├── data/
| ├── external/
| ├── interim/
| ├── processed/
| └── raw/
| ├── Codebook.csv
| ├── Field_Level.csv
| ├── Plot_Level.csv
| └── Soil_PL.csv
|
├── notebooks/
| └── 01_data_audit.ipynb
|
├── outputs/
| ├── figures/
| ├── maps/
| └── tables/
├── report/
| ├── figures/
| └── references/
├── scripts/
├── tests/
├── .gitignore
├── environment.yml
├── LICENSE
├── README.md
├── requirements.txt
```
---
## Project Background
OAF Inc., an agricultural organization based in East Africa, conducted multi-season maize field trials across Rwanda to evaluate the impact of varying urea fertilizer application rates and timing strategies on maize productivity. Data were collected at multiple levels including field, plot, and soil levels across several agroecological regions and mega-environments.
Each field was divided into multiple experimental plots receiving distinct urea rate and timing combinations. Additional management and environmental variables such as planting practices, soil properties, and geographic information were also collected.
The purpose of this analysis is to convert experimental observations into actionable agronomic recommendations and predictive insights that can support evidence-based fertilizer management decisions for farmers and agricultural extension teams.
## Business Problem
Agricultural productivity among smallholder farmers varies substantially across geographic environments due to differences in soil conditions, management practices, and e …