Binary Classification ML Pipeline

The Binary Classification app is a full data-science workflow for binary target columns. Upload a CSV, pick your target, and walk through fifteen pipeline steps in the sidebar.

Data understanding covers profiling, quality scoring, missing values, outliers, and duplicate detection. Preprocessing includes encoding, scaling, and automated feature engineering.

Modeling compares classifiers with cross-validated ROC-AUC, tunes hyperparameters with Optuna, and explains results with SHAP and feature importance charts.

At the end you get business insights and a downloadable executive PDF report for stakeholders.

Stack: Python, Streamlit, pandas, scikit-learn, Optuna, SHAP, Plotly, and ydata-profiling. Deployed on Render with Docker support for local runs.

Try it from the Binary Classification icon on the home page, or reach out at vekris363@gmail.com.