Titanic Survival Predictor
A classic end-to-end ML app that predicts passenger survival, with model explainability built in.
Technical Information
Project Type
ClassificationExplainable MLStreamlit App
Tech Stack
PythonPandasNumPyscikit-learnMatplotlibSeabornJoblibStreamlitHugging Face Spaces
Topics / Modules
Feature EngineeringClassificationModel Explainability
Links
Project Summary
An interactive Streamlit app that predicts Titanic passenger survival from input features, complete with AI explainability to show which factors drove each prediction.
Skills Demonstrated
Feature EngineeringClassificationModel SerialisationExplainable AIStreamlit Deployment
Solution
A scikit-learn classifier trained on the Titanic dataset is serialised with Joblib and served via Streamlit, presenting predictions alongside an explainability breakdown.
Approach
- Prepare: Clean the dataset and engineer features.
- Train: Fit and tune a scikit-learn classifier.
- Explain: Add explainability to surface feature contributions.
- Serve: Build a Streamlit UI for live predictions.
- Deploy: Host on Hugging Face Spaces.