Abdul-Pandev/ghana-agri-decision-support

An agricultural decision support tool that helps Ghanaian farmers choose the right crop, predict yield, and estimate harvest timelines, powered by soil, rainfall, and agronomic data. # HarvestIQ Ghana ### An Agricultural Decision Support Tool for Ghanaian Smallholder Farmers --- ## The Problem Every planting season, thousands of smallholder farmers across Ghana make critical decisions with almost no data. Which crop should I grow? Will my soil support it? How much should I expect to harvest? When should I tell buyers my produce will be ready? These decisions are made by instinct, tradition, and hope. A wrong crop choice, a missed harvest window, or an overestimated yield can mean months of lost income for a family that had no margin to lose. In a country where agriculture employs over 40% of the workforce and the majority of farmers operate below one hectare of land, the cost of a bad decision is not just financial — it is deeply personal. **HarvestIQ Ghana** is a machine learning pipeline built to answer the three questions every Ghanaian farmer, agricultural buyer, and extension officer needs answered before a single seed goes into the ground: > **What should I plant? How much will I get? When will it be ready?** --- ## What This Project Does HarvestIQ Ghana trains two separate regression models on farm-level data from 10 Ghanaian regions: | Model | Target | Question Answered | |---|---|---| | **Yield Predictor** | `Yield_kg_per_ha` | How much will I harvest per hectare? | | **Harvest Timeline Predictor** | `days_to_harvest` | How many days from planting to harvest? | Both models are trained on soil nutrient composition, rainfall, crop type, farm region, and planting season data. The pipeline covers the full ML workflow — raw data → cleaning → feature engineering → EDA → modelling → evaluation. --- ## Who This Is For This tool is designed with three users in mind, and all three benefit from the same pipeline: **The Farmer** — chooses a crop, inputs their soil readings and region, and gets an expected yield and harvest date. Enough information to plan their season, negotiate with buyers, and manage their household income. …