mmsaleem/carbon-membrane-selectivity-app

🎯 CO₂/N₂ Selectivity Predictor – Carbon Membrane App

This Streamlit-based web app predicts CO₂/N₂ selectivity using permeability data of carbon membrane polymers. It's built on a trained machine learning model and is designed for researchers, engineers, and students working on carbon capture technologies.

🚀 Features

✅ Easy-to-use interface ✅ Upload or input new polymer membrane data ✅ Predict CO₂/N₂ selectivity instantly ✅ Visual results & interactive layout ✅ Open-source & publicly accessible

🧪 Model Training

• —Model: RandomForestRegressor
• —Trained using a cleaned membrane dataset with 18 original features
• —Feature engineering includes:
• —One-hot encoding of categorical columns
• —Grouping rare polymers
• —Handling missing values
• —Output: PolyMemCO2Pipeline.joblib (saved ML pipeline)

📦 Requirements

Install dependencies locally:

pip install -r requirements.txt