← Back to Projects

Crop Yield Prediction

A machine learning system to predict crop yields using weather, soil, and historical agricultural data to support farmers and policy planners.

Problem

Crop yields are affected by rainfall, temperature, soil quality, and farming practices. Farmers and governments need accurate yield predictions to plan food supply, pricing, and storage. This project builds a predictive model to estimate crop yield before harvest.

Dataset

The dataset includes historical crop yields, rainfall, temperature, humidity, and soil characteristics across multiple regions and seasons.

Feature Engineering

  • Average rainfall during the growing season
  • Temperature and humidity trends
  • Soil nutrient levels
  • Crop type and seasonality

Model

Regression models such as linear regression, random forest, and gradient boosting were tested. The best-performing model was selected based on prediction accuracy and stability.

Results

The final model was able to predict crop yields with good accuracy, helping identify high-risk regions and seasons where crop failure is more likely.

What I Learned

This project showed how machine learning and environmental data can be combined to support agriculture and food security. It also emphasized the importance of data quality and feature selection.