# Ghana Maize Price Forecasting
A data science project investigating whether the unusually low maize prices seen in Ghana during Q1 2026 represented a structural market shift or a temporary anomaly — motivated by a real conversation with a family member in the maize business.
---
## Background
Maize is a staple commodity in Ghana, with prices following a well-established seasonal cycle: rising during the lean season (peaking around March–April) and falling sharply after the main harvest (September). In Q1 2026 — a period when prices would normally be at their seasonal peak — prices were instead unusually low, prompting the question: *is something fundamentally changing in this market, or is this a transient shock?*
This project uses historical data and machine learning to answer that question.
---
## Project Structure
```
ghana-maize-predictor/
│
├── EDA.ipynb # Exploratory data analysis
├── Forecasting.ipynb # XGBoost forecasting model
├── full_df.csv # Merged dataset output from EDA
│
├── ghana_maize_prices.xlsx # Monthly maize prices (GHS per 100kg)
├── Maize_Production_Ghana.xlsx # Annual maize production (tons)
├── Maize_Imports_-_Ghana.xlsx # Monthly maize imports (tons)
├── Ghana_Inflation_History.xlsx # Monthly YoY inflation (%)
│
└── requirements.txt # Python dependencies
```
---
## Data Sources
| Dataset | Description | Period |
|---|---|---|
| Maize Prices | Monthly price per 100kg bag in Ghana Cedis | 2019 – 2026 |
| Maize Production | Annual domestic production converted to tons | 2019 – 2026 |
| Maize Imports | Monthly import volumes converted to tons | 2019 – 2026 |
| Inflation | Year-on-year inflation percentage | 2019 – 2026 |
The four datasets are merged on Year and Month in the EDA notebook and saved as `full_df.csv` for use in the forecasting notebook.
---
## Methodology
### EDA (`EDA.ipynb`)
- Merges price, production, import, and i …