Project Deep Dive
Flood Prediction ML Model
PythonScikit-learnPandasXGBoost
About the Project
This project predicts flood probability using supervised machine learning on an environmental dataset sourced from Kaggle. The workflow focuses on robust preprocessing, domain-aware feature engineering, and comparative model evaluation for reliable risk estimation.
Feature engineering included crafted interaction terms such as Rainfall × Deforestation to capture compounding environmental effects often missed by linear assumptions.
Evaluation results from the final analysis showed:Random Forest R2: 0.763XGBoost R2: 0.938 (Mean CV: 0.984)
Project Structure
Flood-Prediction-ML-model/ ├── data/ │ └── flood_dataset.csv ├── notebooks/ │ └── model_experiments.ipynb ├── src/ │ ├── preprocess.py │ └── train.py └── requirements.txt
Installation & Usage
Installation
pip install -r requirements.txt
Usage
python src/train.py
Model Performance Visualization
Actual vs predicted comparisons for both models from the final analysis pipeline.
