Hello Streamlit team,
I’m a big fan of Streamlit and use it extensively in my projects. It’s been a while since I built a full-stack bank customer churn predictor using Streamlit, Supabase, Groq API, and EmailJS.
The app analyzes 4000 customer records using models like XGBoost (84.25% accuracy), SVM, Random Forest, and others. Users can input customer details to receive churn risk predictions, view interactive visualizations via Plotly, and get AI-generated explanations. Authentication is persistent with session and state management handled via JSON. Email notifications to me are triggered on signup and password changes using EmailJS. Groq is used for both analysis and generating email drafts.
The app is responsive, supports multiple themes (including transparent mode), and includes a significant amount of custom CSS and JavaScript directly integrated within the Python files.
App: https://cust-churn-pred-bank.streamlit.app/
GitHub: https://github.com/Soumilgit/XYZ-Bank-Customer-Churn-Predictor
It’s also LIVE on Product Hunt !
If possible, I’d love for it to be considered for the Streamlit Gallery.
Thanks for all the amazing work you do.