Rainfall Prediction

•Empowered 20 - 25 farmers by developing a ML system to predict rainfall patterns over 3 days with 75% accuracy
which gave them more insight regarding which crop to choose and when to harvest
• Evaluated Logistic Regression, XGBoost, and Neural Networks to boost the existing system by 12%
• Leveraged 4 unique sensors over 5 fields to analyze 100k data points. Conducted feature engineering with
temperature, dewpoint, and humidity to evaluate feature importance and causal analysis