AI-Powered Water Quality Classifier and Dosage Recommender
This dataset and model release is part of a student-led research project focused on accessible water purification. It includes:
A machine learning classifier that predicts if water is SAFE or UNSAFE based on 10 input features.
Four regression models that recommend corrective dosages of Zeolite A, Alum, Hydrogel, and Titanium Dioxide if the water is unsafe.
Cleaned training dataset (merged_training.csv) with 14,551 samples.
Performance metrics, input schemas, and reproducible notebooks.
This release is made available under an open license to support reproducibility and global water innovation efforts.Compatible with Python 3.10+, scikit-learn, and pandas.
Project supervised and conducted independently as part of Egypt's national delegation to the Stockholm Junior Water Prize 2025.